SYNTHESIS AND CHARACTERIZATON OF INORGANIC MATERIALS FOR SODIUM - ION BATTERIES By Rengarajan Shanmugam A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemical Engineering Doctor of Philosophy 201 5 ABSTRACT SYNTHESIS AND CHARACTERIZATION OF INORGANIC MATERIALS FOR SODIUM - ION BATTERIES By Rengarajan Shanmugam Development of low - cost energy storage devices is critical for wide - scale implementation of intermittent renewable energy technologies and improving the electricity grid. Commercial devices remain prohibitively expensive or lack the performance specifications for a wider market reach. N a - ion batteries would perfectly suited for these large - scale applications as the r aw materials (such as soda ash, salt, etc.) are plentiful, inexpensive and geographically un constrained . However , extensive materials r esearch on insertion electrodes is required for better understanding of the electrochemical and structural properties and engineering high performance Na - ion batteries . This thesis research involves exploratory study on n ew insertion materials w ith various crystallographic structure - typ es and extensive characterization of promising new inorganic compositions . T unnel - type materials , sodium nickel phosphate - Na 4 Ni 7 (PO 4 ) 6 , and sodium cobalt titanate - Na 0.8 Co 0.4 Ti 1.6 O 4 , were investigated to capitalize on the intrinsic structural stability o ffered by framework materials . Sol - gel a nd solid - state reaction synthetic techniques were employed for inorganic powder synthesis. Galvanostatic and potentiostatic testing confirm reversible s odium insertion/de - insertion reactions albeit with inadequate electrochemical characteristics ( high voltage hysteresis> 1V) . Subsequent efforts involved investigating layer - stru ctured materials supporting fast ionic transport for better electrochemical performance . P2 - s odium nickel titanate , Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 (P2 NT ) , with prismatic sodium co - ordination , was synthesized by solid - state technique . The oxide contains Ni 2+/4+ and Ti 4+/3+ redox couples with redox potential s of 3.6 V, 0.7 V vs. Na/Na + , respectively . T h is bifunctional approach would simplify electrode processing and provide cost reduction opportunities in battery m anufacturing . The structural changes monitored using ex - situ XRD demonstrate a favor ably broad solid - solution domain . M anganese substitution, to form P2 - Na 2/3 [Ni 1/3 Mn 1/3 Ti 1/3 ]O 2 (P2NMT) , provides a n enhanc ed high - current performance due to faster interfacial kinetics and accelerated charge carrier transport as shown by impedance spectroscopy and DC testing. Structural properties of P2 NT material were studied using n eutron diffrac ti on and atomisitic simulations . R ie tveld refinement shows that Na f sites have lower site occupancy than Na e sites due to unfavorable repulsive interactions from inline transition metal atom s. Buckingham and Morse - type models accurately predicted the exper imental lattice parameters . The e nergy landscape was explored using energy minimization runs on disordered supercells . T he simulated density maps are in agreement with the experiment densities with evidence of stacking fault formation. O3 - sodium nickel titanate, Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 (O3NT) with octahedral s odium co - ordination was synthesized by solid - state reaction technique. The influence of titanium on the poor cycleability of the O3 - type electrodes w as investigated. Ex - situ XRD shows two phase region s, co mprised of O3+P3 phases , and a solid solution region , comprised of P3 phase . O3NT provides an excellent c apacity retention of 99% for 115 cycles at C/2 rate. The good cycleability is attributed to the relative invariance of net impedance during electr ode cycling using impedance spectroscopy . iv ACKNOWLEDGEMENTS I would like to recognize the valuable research guidance of m y advisor Prof . Wei Lai towards my d octorate degree . I am grateful for the opportunity to do my thesis research under his mentorship and develop my professional skills. I would like to extend my thanks to Prof. Tim Hogan, Prof. Lawrence Drzal, and Prof. Jeff Sakamoto for being a part of my research committe e and providing feedback on my research activities. I would like to thank my lab mates Yuxing Wang, and Matt Klenk. I would like to acknowledge Yuxing for providing in sights on research projects from a materials science perspective and support on lab equi pment and facilities . I t was a pleasure to share lab and office space wit h you guys, and engage in impromptu discussions on various scientific and non - work related topics . I acknowledge CHEMS department for providing multiple teaching assistantships and va lue the lecturing opportunit ies under the mentoring of Prof. Scott Calabrese Barton, Prof. Carl Boehlert, Maddalena Fanelli, Prof. Andre Lee and Prof. Timothy Hinds . I appreciate the research staff at CMSC including Michael Rich and Per Askeland for provi ding training and valuable feedback on electron microscop y , surface area analysis , and electrospinning tools . I also thank the CAM center staff: Stanley Flegler, Carol Flegler, and Abigail Va nderberg for 7500F SEM training, sample preparations and troubles hooting . I acknowledge Prof. Richard Staples for providing training, access and inputs in XRD characterization at chemistry department, MSU. I would like to thank all my friends for the support and making my life in graduate school exciting. Last but not the least, I am grateful for the encouragement, and emotional support of my family . 1 TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ... 5 LIST OF FIGURES ................................ ................................ ................................ . 6 1. INTRODUCTION ................................ ................................ .............................. 14 1.1 Climate Change: A Major Challenge of the 21st Century ................................ ...................... 14 1.2 Sources of GreenHouse Gas Emissions ................................ ................................ .................. 15 1.3 Renewable Energy Technologies: Benefits and Issues ................................ ........................... 16 1.4 Energy Storage Technologies for Stationary Applications ................................ ..................... 19 1.4.1 Pumped Hydroelectric Storage ................................ ................................ ............................ 20 1.4.2 Compressed Air Energy Storage ................................ ................................ .......................... 20 1.4.3 Electrochemical Energy Storage ................................ ................................ .......................... 21 1.4.3.1 Na/S batteries ................................ ................................ ................................ .................... 21 1.4.3.2 Lead - acid batteries ................................ ................................ ................................ ............ 22 1.4.3.3 Li - ion batteries ................................ ................................ ................................ .................. 22 1.4.4 Flywheels ................................ ................................ ................................ ............................. 23 1.5 Market need for stationary sector ................................ ................................ ........................... 24 1.6 Need for Alternate Battery Chemistry ................................ ................................ .................... 25 1.7 Rationale for investigating Na - based bat teries ................................ ................................ ....... 28 1.7.1 Na - ion vs Mg - ion batteries ................................ ................................ ................................ .. 28 1.7.2 Relative ease of scalability ................................ ................................ ................................ ... 29 1.8 Insertion - based Batteries: Working Principle ................................ ................................ ......... 30 1.9 Requisites for Intercalation - based Electrodes ................................ ................................ ......... 31 1.10 Research Problem ................................ ................................ ................................ ................. 33 1.11 Thesis organization ................................ ................................ ................................ ............... 34 BIBLIOGRAPHY ................................ ................................ ................................ ......................... 36 2. SYNTHESIS, AND EL ECTROCHEMICAL CHARAC TERIZATION OF NOVEL TUNNEL - TYPE SODIUM INTERCAL ATION MATERIALS ....... 40 2.1 Introduction ................................ ................................ ................................ ............................. 40 2.2 Structural features ................................ ................................ ................................ ................... 41 2.2. 1 Sodium nickel phosphate: Na 4 Ni 7 (PO 4 ) 6 ................................ ................................ ............. 41 2.2.2 Sodium cobalt titanate: Na 0.8 Co 0.4 Ti 1.6 O 4 ................................ ................................ ............ 42 2.3 Inorganic Synthesis ................................ ................................ ................................ ................. 43 2.3.1 Sodium nickel phosphate synthesis ................................ ................................ ..................... 43 2.3.2 Sodium cobalt titanate synthesis ................................ ................................ .......................... 48 2.4 Film fabrication ................................ ................................ ................................ ....................... 50 2.4.1 Substrate selection ................................ ................................ ................................ ............... 50 2.4.2 Casting methods ................................ ................................ ................................ ................... 50 2 2.5 El ectrochemical characterization ................................ ................................ ............................ 52 2.5.1 Cell setup and testing ................................ ................................ ................................ ........... 52 2.5.2 Evaluation of the electrolyte and current collectors ................................ ............................ 53 2.5.3 Validation of electrochemical methods using standard intercalation materials: Na 0.7 CoO 2 and LiFePO 4 ................................ ................................ ................................ ....................... 59 2.5.4 SNP Electrochemistry : Galvanostatic testing ................................ ................................ ...... 61 2.5.5 SNP Ex - situ XRD characterization ................................ ................................ ...................... 62 2.5.6 SCT Electrochemistry: Potentiostatic Intermittent Titration Testing (PITT) ...................... 64 2.6 Summary ................................ ................................ ................................ ................................ . 66 BIBLIOGRAPHY ................................ ................................ ................................ ......................... 68 3. SYNTHESIS, AN D ELECTROCHEMICAL CH ARACTERIZATION OF P - TYPE LAYER - STRUCTURED MATERIALS : Na 2/3 [Ni 1/3 Mn x Ti 2/3 - x ]O 2 (x=0, 1/3) ................................ ................................ ................................ .................. 72 3.1 Introduction ................................ ................................ ................................ ............................. 72 3.1.1 Ionic conduction in layered materials ................................ ................................ .................. 72 3.1.2 Literature reports on layered host materials ................................ ................................ ......... 73 3.1.3 Materials of interest ................................ ................................ ................................ ............. 74 3.2 Structural features ................................ ................................ ................................ ................... 75 3.2. 1 Sodium nickel titanate: P2 - Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 ................................ ................................ ..... 75 3.2.2 Sodium nickel manganese titanate: P2 - Na 2/3 [Ni 1/3 Mn 1/3 Ti 1/3 ]O 2 ................................ ......... 76 3.3 Inorganic synthesis and high - temperature processing ................................ ............................ 76 3.3.1 Sodium nickel titanate synthesis ................................ ................................ .......................... 76 3.3.2 Sodium nickel manganese titanate synthesis ................................ ................................ ....... 79 3.3.3 Powder sintering ................................ ................................ ................................ .................. 81 3.4 Film fabrication ................................ ................................ ................................ ....................... 82 3.5 Material Characterization ................................ ................................ ................................ ........ 82 3.5.1 Voltage profile under galvanostatic testing ................................ ................................ ......... 82 3.5.1.1High - voltage cathode: Ni 2+/4+ redox couple ................................ ................................ ....... 82 3.5.1.2 Low - voltage anode: Ti 4+/3+ redox couple ................................ ................................ .......... 84 3.5.2 Quantifying charge carrier transport processes: DC & AC Testing ................................ .... 86 3.5.3 Evaluating ionic transport and interfacial processes: Electrochemical impedance spectroscopy ................................ ................................ ................................ ....................... 90 3.5.4 Analyzing ionic transport processes: Potentiostatic intermittent titration technique ........... 97 3.5.5 Understanding re action mechanisms: Ex - situ XRD Testing ................................ ............. 100 3.5.6 Evaluating capacity retention ................................ ................................ ............................. 102 3.5.6.1 Cycling Study ................................ ................................ ................................ .................. 102 3.5.6.2 Impedance change during cycling ................................ ................................ .................. 104 3.5.7 Rate performance testing ................................ ................................ ................................ ... 106 3.6 Improving cathodic stability: Calcium substitution ................................ .............................. 108 3.7 Improving anodic capacity: Potassium substitution ................................ ............................. 109 3.8 Summary and future work ................................ ................................ ................................ .... 111 BIBLIOGRAPHY ................................ ................................ ................................ ....................... 114 3 4. STRUCTURAL CHARAC TERIZATION OF P2 - SODIUM NICKEL TITANATE USING EXPER IMENTAL AND COMPUTAT IONAL METHODS ................................ ................................ ................................ ........... 118 4.1 Introduction ................................ ................................ ................................ ........................... 118 4.1.1 Comp utational methods ................................ ................................ ................................ ..... 118 4.1.2 Literature ................................ ................................ ................................ ............................ 119 4.1.3 Scope of current work ................................ ................................ ................................ ........ 119 4.2 Experimental ................................ ................................ ................................ ......................... 120 4.2.1 Neutron diffraction experiments ................................ ................................ ........................ 120 4.2.2 Energy calculations for many - body systems: Theory ................................ ........................ 120 4.2.3 Long - range and short - range interactions ................................ ................................ ........... 121 4.2.4 Handling partial occupancy: Mean field model and Supercells ................................ ........ 125 4.2.5 Energy minimization: Optimization schemes and strategy ................................ ............... 126 4.2.6 Space group transformations ................................ ................................ .............................. 127 4.2.7 Simulation data interpretation and Pair Distribution Function (PDF) calculation ............ 130 4.3 Results and Discussion ................................ ................................ ................................ ......... 131 4.3.1 Rietveld refinement: Neutron Diffraction ................................ ................................ .......... 131 4.3.2 Interatomic potential sets: Selection and validation ................................ .......................... 136 4.3.3 Energetics of disordered structures: Ato mic distribution effects ................................ ....... 140 4.3.4 Energetics of ordered structures ................................ ................................ ......................... 144 4.3.5 Simulated average structure ................................ ................................ ............................... 145 4.3.6 Pair distribution function and probability density line - scans ................................ ............ 150 4.4 Summary and future work ................................ ................................ ................................ .... 152 BIBLIOGRAPHY ................................ ................................ ................................ ....................... 154 5. SYNTHESIS, STRUCT URAL AND ELECTROCHEM ICAL CHARACTERIZATION OF Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 , AN O - TYPE LAYER - STRUCTURED MATERIAL ................................ ................................ ............. 159 5.1 Introduction ................................ ................................ ................................ ........................... 159 5.1.1 O3 - type materials ................................ ................................ ................................ ............... 159 5.1.2 Literature ................................ ................................ ................................ ............................ 160 5.1.3 Rese arch Strategy ................................ ................................ ................................ ............... 161 5.2 Material synthesis & electrode fabrication ................................ ................................ ........... 162 5.3 Electrochemistry ................................ ................................ ................................ ................... 164 5.3.1 Voltage - composition curves: Galvanostatic Testing ................................ ......................... 164 5.3.2 Phase transitions: Ex - situ XRD ................................ ................................ ......................... 165 5.3.3 Equilibrium voltage profiles and solid - state diffusivi ty estimation: Potentiostatic Intermittent Titration Technique (PITT) ................................ ................................ .......... 170 5.3.4 Cycling study ................................ ................................ ................................ ..................... 172 5.3.4.1 Capacity retention at different voltage cutoffs ................................ ................................ 172 5.3.4.2 Impedance spectra evolution during cycling ................................ ................................ .. 175 5.3.4.3 Sodium metal deactivation: Prolonged cycling ................................ .............................. 178 5.3.5 Rate testing ................................ ................................ ................................ ......................... 179 5.4 Summary and future work ................................ ................................ ................................ .... 180 BIBLIOGRAPHY ................................ ................................ ................................ ....................... 183 4 6. CONCLUSIONS ................................ ................................ .............................. 186 5 LIST OF TABLES Table 1.1 Short - term and Long - term DOE targets for EES technologies (stationary applications). 25 ................................ ................................ ................................ ................................ ....................... 25 Table 1.2 Comparison of resources and physical properties of lithium, sodium and magnesium 30 ................................ ................................ ................................ ................................ ....................... 27 Table 3.1 Theoretical capacity of P2 - titanate materials based on vacancy concentration. ......... 110 Table 4.1 Interatomic potential parameters: Buckingham Core - shell model; all atomic species are taken at formal charges, Formal charge=Y core + Y shell ................................ ................................ . 124 Table 4.2 Interatomic potential parameters: Morse+ LJ model ................................ .................. 124 Table 4.3 Structural information of P2 - NT powders: Prepared at 900 o C and tested at 15K. .... 133 Table 4.4 Structural information of P2 - NT powders: Prepared at 900 o C and tested at 300K ... 134 Table 4.5 Structural information of P2 - NT powders: Prepared at 1000 o C and tested at 15K ... 134 Table 4.6 Structural information of P2 - NT powders: Prepared at 1000 o C a nd tested at 300K . 134 Table 4.7 modulus percentage error in lattice parameters. ................................ ................................ ......... 139 Table 4.8 Transferability of Buckingham interatomic potential across ternary oxides. ............. 139 Table 4.9 Transferability of Morse interatomic potential across binary oxides. ........................ 139 Table 4.10 Transferability of Morse interatomic potential across ternary oxides. ..................... 139 6 LIST OF FIGURES Figure 1.1 (a) Global carbon dioxide uptake over past decades, (b) Surface temperature anomalies; 3 For interpretation of references to color in this and all other figures, the reader is referred to the electronic version of this thesis. ................................ ................................ ................................ .... 15 Figure 1.2 (a) Global greenhouse gas emissions, 12 and (b) Gl obal energy consumption in past decades. 13 ................................ ................................ ................................ ................................ ...... 16 Figure 1.3 (a) Global energy generation by sources,17 (b) Renewable energy estimated output and projections. 17 ................................ ................................ ................................ ................................ . 17 Figure 1.4 Variability in (a) solar power output and, (b) wind power output. 18 ........................... 18 Figure 1.5 Schematic representation of (a) pumped hydroelectric system, (b) compressed air energy storage system, (c) electroc hemical energy storage system with Na/S battery chemistry, and (d) flywheel system. 22 ................................ ................................ ................................ ............ 21 Figure 1.6 (a) Cost and efficiency compa rison of different energy storage technologies; hatched bars refer to the efficiency values on the secondary y - scale , and (b) lifetime comparison of energy storage technologies. 21 ................................ ................................ ................................ .................. 22 Figure 1.7 Market segments for energy storage devices in the electricity grid. ........................... 24 Figure 1.8 Energy storage technologies employed for various stationary applications. 21 ............ 24 Figure 1.9 Relative abundance chart depicting plentiful sodium element as compared to lithium. 29 ................................ ................................ ................................ ................................ ....................... 27 Figure 1.10 Manufacturing steps for advanced batteries such as Li - ion chemistry. ..................... 29 Figure 1.11 Workin g principle of an insertion/intercalation - based battery. 34 .............................. 31 Figure 1.12 Growing number of research publications on anode materials for Na - ion battery. 36 34 7 Figure 2.1 Crystal structure of Na 4 Ni 7 (PO 4 ) 6 ; Pink polyhedra represents nickel octahedra, gree n polyhedra represents phosphate tetrahedra and yellow spheres indicate sodium atoms. Oxygen atoms are not included for clarity. 26 ................................ ................................ ............................. 42 Figure 2.2 Crystal structure of Na 0.8 Co 0.4 Ti 1.6 O 4 ; blue polyhedra represents transition metal octahedra (cobalt and titanium) and yellow spheres represent sodium atoms; oxygen atoms are omitted for clarity. 25 ................................ ................................ ................................ ...................... 43 Figure 2.3 Schematic of solid state reaction technique. ................................ ................................ 44 Figure 2.4 XRD characterization of sodium nickel phosphate prepared by solid state reaction. . 45 Figure 2.5 (a) Sodium nickel phosphate powder morphology prepared by solid state reaction, and (b) Higher - magnification image of a particle showing dense morphology. ................................ . 46 Figure 2.6 Schematic of sol - gel synthesis technique. ................................ ................................ ... 47 Figure 2.7 XRD characterization of sol - g el synthesized powders of sodium nickel phosphate. .. 47 Figure 2.8 (a) Sodium nickel phosphate powder morphology prepared by sol - gel reaction, and (b) Higher - magnification image of a particle showing porous morphology. ................................ ..... 48 Figure 2.9 (a) XRD characterization of s odium cobalt titanate prepared by solid state reaction, and (b) SEM morphology of as - synthesized sodium cobalt titanate powders. ................................ .... 49 Figure 2.10 Schematic of doctor - blade slurry casting method for fabricating battery electrodes. 51 Figure 2.11 Schematic of Swagelok 3 - electrode cell setup for electrochemical testing. ............. 52 Figure 2.12 Impedance spectroscopy curves for sodium metal SEI formation in (a) 1M NaClO4 in ethylene carbonate/dimethyl carbonate, and (b) 0.5M NaPF6 in ethylene carbonate/diethyl carbonate; 1st cy cle refers to test after 4 hours open - circuit and subsequent tests. ...................... 54 Figure 2.13 Cyclic voltammograms scanned between 1.5 - 4.2 V voltage window for (a) aluminum, (b) andoized aluminum, (c) stainless steel 316, and (d) titanium. ................................ ................ 57 8 Figure 2.14 (a) Cyclic vol tammogram of graphite scanned between 1.5 - 4.2 V voltage window, and (b) oxidative voltammogram peak currents for 1 st and 10 th cycle. ................................ ............... 58 Figure 2.15 Voltage profile of anodized aluminum current collector tested under galvanostatic condition with 4.5V cutoff using 3 - electrode setup. ................................ ................................ ..... 59 Figure 2.16 Voltage profiles of (a) lithium iron phosphate under galvanostatic conditions at C/10 rate demonstrating two - phase reaction be tween FePO 4 and LiFePO 4 , and (b) Na 0.7 CoO2 showing complex voltage steps and plateaus due to multiple phase transitions. ................................ ........ 60 Figure 2.17 Sodium nickel phosphate tested under galvanostatic conditions with discharge cut - off voltage of (a) 1.5 V in 1st cycle, and (b) 0.8 V. ................................ ................................ ............ 62 Figure 2.18 Ex - situ XRD characterization of SNP electrode film at different state - of - charge, Mylar film and wax (used for sticking the electrode to the holder base). ................................ ............... 63 Figure 2.19 (a) Voltage curve of sodium cobalt titanate from PITT with C/50 current limit, 100 mV step potential tested at 70 o C using 0.5M NaPF6 in EC/DEC electrolyte, and (b) Discharge capacity and faradaic efficiency with cycle number. ................................ ................................ .... 65 Figure 2.20 (a) Voltage profile of sodium cobalt titanate tested under at C/100 rate usi ng 1M NaClO4 in EC/DMC electrolyte at room temperature, and (b) Ex - situ XRD of SCT at pristine, 20% charged and discharged state. ................................ ................................ ............................... 66 Figure 3.1 Schematic of the transition metal slab in layered oxide depicting charge carrier pathways in (a) P2 - NT, and (b) P2 - NMT through formation of percolating network; red, blue and orange spheres denote nickel, titanium, and manganese atoms, respe ctively. ............................. 75 Figure 3.2 Crystal structure of Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 ; blue spheres refer to transition metals (nickel and titanium), yellow spheres refer to sodium atoms Oxygen atoms are not included for clarity. ................................ ................................ ................................ ................................ ....................... 76 Figure 3.3 XRD characterization of sodium nick el titanate prepared by solid state reaction from micron - sized precursors fired for 12 hours duration; The dominant peaks of the impurity and target phases are indexed. ................................ ................................ ................................ ....................... 77 Figure 3.4 XRD characterization of P2 - NT material prepared by solid state reaction from wet - milled nano - sized precursors fired at (a) 800 o C, and (b) 900 o C, SEM morphology of powders prepared from (c) micron - sized precursors, and (d) nano - sized precursors. ................................ 79 9 Figure 3.5 (a) XRD comparisons of sodium nickel titanate and sodium nickel mangane se titanate powders prepared by solid state reaction, (b) and SEM powder morphology. ............................. 80 Figure 3.6 (a) Surface morphology of sintered pellets of (a) P2 - Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 , and (b) P2 - Na 2/3 [Ni 1/3 Mn 1/3 Ti 2/3 ]O 2 . ................................ ................................ ................................ ............... 81 Figure 3.7 Cathodic voltage profile under galvanostatic testing rate of C/50 rate for (a) P2 - NT material with cutoff of 4.2 V and, (b) P2 - NT material with cutoff of 4.5 V, (c) P2 - NMT material with cutoff of 4.2 V and, (d) P2 - NMT material with cutoff of 4.5 V. ................................ .......... 83 Figure 3.8 Anodic voltage profile of sodium nickel titanate under galvanostatic testing of C/5 rate with 0.2 V cutoff voltage. ................................ ................................ ................................ ............. 85 Figure 3.9 (a) Electrical equivalent circuit model representing physical processes in a mixed electronic - ionic conductors with blocking electrodes, (b) Limiting model for high - frequency AC signal, and (c) Limiting model for DC signal. 28 ................................ ................................ ........... 87 Figure 3.10 Electrical testing of P2 - NMT s intered pellet: (a) High - frequency AC impedance spectra in temperature range of 383 - 423 K with 10 K increments, and (b) Current decay response upon applying a potential of 200 mV at a 383 K. ................................ ................................ ......... 88 Figure 3.11 Total effective conductivity (ionic and electronic) and electronic conductivity curves as a function of temperature from AC impedance and DC testing. ................................ .............. 89 Figure 3.12 Electrical equivalent circuit model depicting sodium ion transport processes (within liquid electrolyte, across surface films, across the interf ace, and bulk solid state diffusion) of the composite electrode; where Re is the electrolyte resistance, Rsf is the surface film resistance, CPE sf is the constant phase element due to surface films, CPE ct is the constant phase element due to the double lay er, and Z w is the semi - infinite Warburg diffusional element. ................................ ...... 91 Figure 3.13 Nyquist plot of P2 - NT (blue) and P2 - NMT (yellow) electrodes scanned between frequency range of 5 MHz to 10 mHz with inset showing high - frequency d epressed semicircles at voltage of (a) 3.4 V vs. Na/Na + , (b) 3.8 V vs. Na/Na + , and (c) 4.2 V vs. Na/Na + . ....................... 93 Figure 3.14 Potential dependence of (a) charge transfer resistance and double layer capacitance, and (b) surface film resistance and surface film capacitances. ................................ ..................... 94 - 0.5 curves using planar diffusion model to obtain Warburg prefactor. ................................ ................................ ................................ ................................ ....... 96 10 Figure 3.16 Voltage profile under potentiostatic intermittent titration technique using C/100 current limitation for (a) cathodic P2 - NT with 5 mV voltage step, and (b) cathodic P2 - NMT with 5 mV voltage step, and (c) anodic P2 - NT with 15 mV voltage step; in set in cathodic curves depict electrode precycling at C/10 rate. ................................ ................................ ................................ . 98 Figure 3.17 (a) Linear fitting of current transients from pote ntiostatic intermittent titration technique at different electrode potentials with current limitation of C/100 for cathodic P2 - NT electrode, (b) linear fitting of P2 - NMT electrode under identical conditions, and (c) variation of apparent chemical diffusion co - efficient with electrode potential as calculated from potentiostatic intermittent titration technique and impedance spectroscopy. ................................ ...................... 99 Figure 3.18 Ex - situ XRD of electrode films: (a) sodium nickel titanate cathode charged till 4.2 V with inset showing the evolution of 16o reflection, (b) sodium nickel titanate cathode charged to 4.5 V, (c) sodium nickel manganese titanate cathode charged till 4.2 V, and (d) sodium nickel titanate anode charged to 0.2 V with inset showing evolution of 16 o reflection; all potentials are measured with respect to Na/Na + . ................................ ................................ ............................... 101 Figure 3.19 Capacity retention study with 4.2 V upper voltage cutoff for (a) cathodic sodium nickel titanate at C/10 rate, (b) cathodic sodium nickel manganese titanate electrodes tested at C/5 rate, and (c) anodic sodium nickel titanate tested at C/5 rate. ................................ ............................ 104 Figure 3.20 (a) Nyquist impedance spectroscopy plot of sodium nickel manganese titanate electrode scanned in frequency range of 5 MHz to 10 mHz after charging to 4.2 V, (b) variation of impedance resistive parameters, and (c) capacitive parameters during cycling. ........................ 105 Figure 3.21 High - current rate testing of (a) cathodic sodium nickel titanate powders with micron - and nano - sized powders, (b) cathodic sodium nickel titanate and sodium nickel manganese titanate electrodes with 4.2 V cutoff, and (c) anodic sodium nickel titanate with cutoff voltage of 0.2 V. ................................ ................................ ................................ ................................ ..................... 107 Figure 3.22 (a) XR D characterization of 5% and 20% calcium substituted sodium nickel titanate prepared by solid - state reaction at 900 o C, and (b) EDS mapping of 5% sodium nickel titanate powders to check calcium distribution. ................................ ................................ ...................... 108 Figure 3.23 XRD characterization of powders to synthesize (a) sodium deficient compositions with x=0.67, 0.55 and 0.50 at firing temperature of 1000 o C for 20 h duration, and (b) Na 2/3 - y K y [Li 1/3 Ti 2/3 ]O 2 for y=0, 0.05, 0.10 and 0.15 prepared at firing temperature of 800 o C for 20 h. ... 110 Figure 4.1 Variation of interaction energy for Buckingham and Morse interatomic potentials for Na - O two - body interactions. ................................ ................................ ................................ ....... 125 11 Figure 4.2 Unit cell transformation from space group 194 to 63. ................................ ............... 128 Figure 4.3 Space group transformation to generate in - plane ordered structure. ......................... 129 Figure 4.4 Space group transformation to generate th rough - plane ordered structure. ............... 130 Figure 4.5 Rietveld refinement of P2 - NT po wders prepared at 900 o C and characterized at (a) 15K, bank 1, (b) 15K, bank 4, (c) 300K, bank 1, and (d) 300K, bank 4; red point indicates experimental data, green curve indicates simulated structure, and pink curve indicates difference curve; bank 1, ban k 2 includes diffraction data in the d - spacing range of 0 - 2.2 Å, and 1.1 - 8 Å, respectively. . 132 Figure 4.6 Rietveld refinement of P2 - NT powders prepared at 1000 o C and characterized at (a) 15K, bank 1, (b) 15K, bank 4, (c) 300K, bank 1, and (d) 300K, bank 4. ................................ ... 133 Figure 4.7 3D nuclear density map using Rietveld refined 900 o C, 15K structure (a), {100} planar density describing sodium (b), transition metal (c) and oxygen atoms (d) with minimum and maximum saturation levels at 0 and 5 Å - 3 ; atomic species legend is included on the right - side of 3D density map. ................................ ................................ ................................ .......................... 136 Figure 4.8 Crystal structure of (a) cubic sodium o xide (Na 2 O) , (b) cubic nickel oxide (NiO), (c) tetragonal titanium oxide (TiO 2 ), (d) rhombohedral nickel titanium oxide (NiTiO 3 ), (e) monoclinic sodium titanate - I (Na 2 Ti 3 O 7 ), and (f) monoclinic sodium titanate - II (Na 2 Ti 6 O 13 ); yellow, grey, red, and blue s pheres represent sodium, nickel, oxygen, and titanium, respectively. colored polyhedra represent transition metals co - ordinated with oxygen atoms, sodium are left unbonded for clarity. 40 - 45 ................................ ................................ ................................ ................................ ... 138 Figure 4.9 (a) Buckingham potential energy distribution curve, (b) Morse potential energy distribution curve, (c) probability distribution histogram of randomized structures with equal atomic distribution between layers using Buckingham potential, and (d) probability distribution histogram of randomized structures with equal atomic distribution between layers using Morse potential. ................................ ................................ ................................ ................................ ...... 142 Figure 4.10 Probability distribution histogram for (a) unequal sodium distribution using Buckingham potential, (b) unequal sodium distribution using Morse potential, (c) unequal transition metal distribution using Buckingham potential, and (d) unequal transition metal distribution using Morse potential. ................................ ................................ ............................. 143 Figure 4 .11 Probability distribution histogram for completely randomized structures using (a) Buckingham potential, and (b) Morse potential. ................................ ................................ ........ 144 12 Figure 4.12 Energy distribution curve for ordered and disordered structures using Morse potential. ................................ ................................ ................................ ................................ ..................... 145 Figure 4.13 Simulated sodium probability density maps in (a) 3D using Buckingham potential and (b) 3D using Morse potential, (c) 2D along using Buckingham potential, and (d) 2D using Morse potential; selected 2D cut - sections are parallel to {100} planes; minimum and m aximum saturation levels set at 0 and 5 Å - 3 . ................................ ................................ ................................ .............. 147 Figure 4.14 Simulated probability density maps for (a) sodium atoms in 3D using Buckingham potential and (b) sodium atoms in 3D using Morse potential, (c) sodium atoms in 2D using Buckingham potential, (d) sodium atoms in 2D using Morse potential; 2D cut - sections are parallel to {100} planes; minimum and maximum saturation levels set at 0 and 5 Å - 3 . ......................... 148 Figure 4.15 2D oxygen density maps using (a) Buckingham potential, and (b) Morse potential; 2D cut - sections ar e parallel to {100} planes; minimum and maximum saturation levels set at 0 and 5 Å - 3 . ................................ ................................ ................................ ................................ .............. 149 Figure 4.16 (a) Experimental pair distribution function using Buckingham potential, (b) experimental pair distribution function using Morse potential, (c) normalized probability density between Na f and Na e normalized p robability density between four different Na f - Na e locations using Buckingham potential. ................................ ................................ ................................ ................................ ...... 1 51 Figure 5.1 (a) Rietveld refinement of O3NT powders prepared by solid - state reaction at 1000 o C for 10 hours firing duration with inset figure showing powder morphology, (b) SEM morphology of composite film electrode showing O3NT particles embedded in a carbon network formed by conductive addit ive, and (c) EDS mapping of powders showing elemental distribution of carbon and titanium; refinement was performed with lab powder diffraction dataset using TOPAS software by Bruker Corporation. 13 16 ................................ ................................ .......................... 163 Figure 5.2 Voltage - composition curve of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 under galvanostatic conditions st charging cycle is due to side reactions. ................................ ................................ ................................ ...... 164 Figure 5.3 Crystal structure of (a) O3 phase with octahedral co - ordination of sodium, and (b) P3 phase with prismatic co - ordination of sodium demonstrating the different oxygen stacking sequence. ................................ ................................ ................................ ................................ ..... 166 Figure 5.4 Ex - to the unidentified phase formed due to moisture in tercalation of the low - sodium content phases. ................................ ................................ ................................ ................................ ..................... 167 13 Figure 5.5 Rietveld refinement of O3NT electrode (Na x [Ni 0,45 Ti 0.55 ]O 2 ) prepared by TOPAS software from Bruker Corporation, insert figures depict sodium co - ordination environment and Na - O bond lengths. ................................ ................................ ......................... 169 Figure 5.6 (a) Voltage profile of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 using Potentiostatic Intermittent Titration Technique (PITT) with C/100 current and 4.2 V voltage limit , (b) Voltage profile using PITT with C/100 current and 4.5 V voltage limit, (c) current transient linear fitting with 4.2 V limit, and (d) current transient linear fitting with 4.5 V limit. ................................ ................................ .......... 171 Figure 5.7 Solid - state diffusion co - efficient of sodium ions in O3NT material using potentiostatic intermittent titration testing. ................................ ................................ ................................ ........ 172 Figure 5.8 Capacity study of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 material: (a) 4.0 V charging limit, (b) 4.2 V charging limit, (c) 4.4 V charging limit, and (d) Variation of charging capacity with different cut - off voltage limit. ................................ ................................ ................................ .......................... 174 Figure 5.9 (a) Nyquist plot of O3NT electrodes charged to 4.2 V during cycling, (b) Bode plots of O3NT electrodes charged to 4.2 V during cycling, (c) Variation of electrolyte resistance (R e ), surface film resistance s (Rsf1 and Rsf2), and charge transfer resistance (Rct) during cycling, and (d) Variation of double layer capacitance (Cdl) and surface film capacitances (Csf1 and Csf2) . ................................ ................................ ................................ ................................ ..................... 177 Figure 5.10 Variation of peak counter electrode potential with cycle number during galvanostatic cycling measured against Na/Na + reference electrode; ideally the measured voltage would be zero at the counter el ectrode as it has the same chemical potential as the reference electrode. ......... 178 Figure 5.11 Rate performance testing of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 material with 4.2 V cut - off voltage (a) capacity retention, and (b) voltage profiles. ................................ ............................. 180 14 1. INTRODUCTION 1.1 Climate Change: A Major Challenge of t he 21st Century The amount of atmospheric carbon dioxide has substantially increased over the past two decades and was estimated to be at 395 ppm in 2014 ( Figure 1 . 1 a ). 1 , 2 The increasing amount o f carbon dioxide along with other anthropogenically produced greenhouse gases (methane, nitrous oxide, fluorinated gases, etc.) has led to the warming of the climate system. Global c limate change is anticipated to result in a number of deleterious effects such as higher surface temperature s - land and oceanic ( Figure 1 . 1 b ) 3 , elevation of mean sea level, acidification of the ocean and alteration , natural weather patterns (including precipitation, drought , etc.) . 1 , 3 - 6 Even though these harmful effects are widely recognized, there are significant hurdles to reduce gas emissions due to the complex interplay betw een climate change , ecology, and economy . 7 - 11 International agencies such as United Nations Framework Convention on Climate Change (UNFCC) and Intergovernmental Panel on Climate Change (IPCC) have been striving to achieve a global collective consensus on emission reductions. Climate change has become one of the prime challenges of the 21 st century and advancing and developing alternate green technology solutions is of paramount importance. 15 (a) (b) Figure 1 . 1 (a) Global carbon dioxide uptake over pa st decades, (b) Surface temperature anomalies; 3 For interpretation of references to color in this and all other figures, the reader is referred to the electronic version of this thesis . 1.2 Sources of GreenHouse Gas Emissions The combustion of fossil fuels (coal, natural gas and oil) for energy generation is the largest contributor (26%) of global (Greenhouse gases) GHG emissions ( Figure 1 . 2 a ) based on estimates by U.S. Environmental Protection Agency (EPA). 12 The global energy usage has grown substantially as shown in Figure 1 . 2 b and the currently at 530 Quadrillion Btu in 2012 based on estimates by U.S. Energy Information Administration (EIA). 13 In the light of ever - growing energy needs in the modern society, developing improved low - carbon renewable sources (wind and solar) and wide - scale integration of these technologies are vital for generati ng clean and sustainable energy. The unavailability of cheap, high - performance energy storage devices is considered a major setback for increasing the adoption of renewable energy sources. 14 - 16 16 (a) (b) Figure 1 . 2 (a) Global greenhouse gas emissions, 12 and (b) Global energy consumption in past decades. 13 EPA reports transportation sector to be a significant emissions contributor at 13% ( Figure 1 . 2 ). 12 Emission reductions in this s ector are addressed by government regulations stipulating increased automobile fuel efficiencies spurring the advancement of hybrid and fully electric drivetrains that incorporate energy storage devices (batteries). 1.3 Renewable Energy Technologies: Benefit s and Issues Renewable energy sources such as wind, solar, geothermal, etc. are extensively developed as viable alternative sources of clean energy with minimal carbon footprint. Widespread deployment of renewable energy technologies would provide the foll owing benefits: i. Achieve compliance with emission standards to combat global climate change provides capacity addition to meet the ever - increasing energy demands of the society 17 ii. Reduce dependence on depleting non - renewable fossil fuel reserves and provide economic stability by locally producing energy and cut down on oil imports iii. Provide a cleaner environment without air or water pollution by zero - emissions operation iv. Decentral ize electricity grids and potentially aid in creating self - sufficient/off - grid homes and communities Despite these advantages, renewables accounted for a paltry 19% of the global energy production in 2012 ( Figure 1 . 3 a ). 17 Nevertheless, the electricity output from renewables has shown a promising r ising trajectory ( Figure 1 . 3 b ), and is anticipated to increase further. 17 (a) (b) Figure 1 . 3 (a) Global energy generation by sources,17 (b) Renewable energy estimated output and projections. 17 Renewable technologies, such as solar and wind, are highly intermittent and this poses challenges while catering to the complex electricity demands of the end - consumers . For instance, solar energy has significant variations in the electricity output depending on factors like day/night 18 cycle, cloud cover , sun positioning, etc. as shown in Figure 1 . 4 a . 18 Temporally and geographically uneven wind flow patterns can lead to unpredictable power output from wind energy technology ( Figure 1 . 4 b ). 18 For instance, Germany, with large investments in renewable sources (as a power generat ion further substantiating the intermittency issue. 19 (a) (b) Figure 1 . 4 Variability in (a) solar power output and, (b) wind power output. 18 19 Energy storage devices can act as a buffer by storing excess energy during times of surplus - - cost, long - life energy storage devices woul d help reduce the variability in renewable throughput and improve overall reliability. In addition to being vital cogs for renewables technologies, large - scale energy storage devices can be beneficial to the conventional electricity grid by i. Providing load leveling for traditional power plants (coal, natural gas and nuclear) and improve the overall efficiency by permitting optimal constant capacity operation (no ramp - up or ramp - ii. Enhancing the power quality by voltage and frequency regulation iii. Improving grid reliability by reducing power outages and acting as a backup system 1.4 Energy Storage Technologies for Stationary Applications Different energy storage technologies have been developed and commercialized for stationary applications and includes i. Pumped Hydroelectric Storage (PHS) ii. Compre ssed Air Energy Storage (CAES) iii. Electroch emical Energy Storage (EES) iv. F lywheels (FW) While PHS, CAES and FW technologies harness mechanical conversion processes, EES utilizes chemical conversion methods. A schematic of working mechanisms of these storage technol ogies is illustrated in Figure 1 . 5 . 20 1.4.1 Pumped Hydroelectric Storage PHS operates by pumping water to an elevated reservoir (from ground level) and running it through a turbine to extract the stored potential energy ( Figure 1 . 5 a ). PHS is currently the largest energy storage method and provides 99% of the net global capacity (127 GW). This is because PHS provides a means for large capacity installations at low operational costs ( Figure 1 . 6 a ). 20 The stored energy scales linearly with mass (or volume) of stored water and elevation difference (E P P 13,000 cycles) as depicted in Figure 1 . 6 b. 21 However, constructing new PHS facilities requires specific geographical locations such as hilly terrai ns and accessible large water bodies. Hence, hydroelectric technology is mostly constructed near river canyons with proper topographical features and rock formations. Consequently, it is becomes geographically constrained and suitable only for large - capaci ty installations. 1.4.2 Compressed Air Energy Storage CAES operates by compressing gasses into enclosed spaces (caverns, special rock formations, abandoned oil mines, etc.) as illustrated in Figure 1 . 5 b. It has a net installed capacity of 440 MW, three orders of magnitude lower than PHS. CAES provides comparable lifetime (13,000 cycles) to PHS. However, like PHS this technology also requires specific geographic formations ( to avoid pressure losses to surroundings). Another approach has been make use of high - pressure tanks but this limits the scale of the application due to technological limitations and increases installation costs. 20 21 (a) (b) (c) (d) Figure 1 . 5 Schematic representation of (a) pumped hydroelectric system, (b) compressed air energy storage system, (c) electrochemical energy storage system with Na/S battery chemistry, and (d) flywheel system. 22 1.4.3 Electrochemical Energy Storage 1.4.3.1 Na/S batteries Na/S batteries (EES) operate with molten sodium and sulfur as the active components using an electrochemical setup as illustrated in Figure 1 . 5 c. The modular design allows for easy installation anywhere using scalable battery packs. However, these devices have a shorter lifetime (4500 cycles) and higher installation costs ($535/kWh) than PHS and CAES as shown in Figure 1 . 6 a and b. 21 The high temperature operation using super reactive molten elements poses challenges in 22 material selection and engineering fail - safe mechanisms. In 2011, NAS batteries designed by The Tokyo Electric Power Company caught fire due to leakage issues. (a) (b) Figure 1 . 6 (a) Cost and efficiency compar ison of different energy storage technologies ; hatched bars refer to the efficiency values on the secondary y - scale , and (b) lifetime comparison of energy storage technologies. 21 1.4.3.2 Lead - acid batteries L ead A cid B atteries (LABs) offer s a relatively low - cost EES solution ($450/kWh) using a well - established battery technology and largely used for small - scale backup for homes and businesses as Uninterrupted Power Supply (UPS) devices. However, they have a short cycle life (2200 cycles) whic h adds to the routine replacement costs for applications requiring long lifetime. 1.4.3.3 Li - ion batteries Li - ion batteries (LIBs) offer high - energy density as they are based on highly electropositive and light weight lithium as an active component. LIB demonstrations have been made for relatively 23 small - scale frequency regulation and renewables applications so far with good efficiency (90%). However, they remain fairly expensive for bulk energy storage. Safety issues due to lithium dendritic plating, electrolyte decomposition due to overcharging, etc. can arise if used beyond the recommended operating specifications and in - built circuits coupled with sensors actively monitor individual cell voltages (within a pack). These circuits help avoid overchargi ng, and temperature abnormalities that can potentially lead to battery failures. 23 It is anticipated that high - volume manufacturing of larger LIB packs for automotive drivetrains would lead to cost - reductions and enable increased usage in stationary sectors. 1.4.4 Flywheels Flywheels work by converting electrical energy into kinetic energy in the form of angular momentum of a spinning rotor as illustrated in Figure 1 . 5 d. At a smaller sc ale, they are extensively employed in automobiles to transfer the energy from the power stroke to idle strokes of a combustion cycle and help level - out the power output. Stationary applications require high operational efficiency and this is typically atta ined by spinning the rotors inside low - pressure chambers to minimize frictional losses. Since kinetic energy scales non - linearly with angular velocity (E k =1/2*I*v 2 k velocity), it is beneficial to run these devices at high rotational speeds. However, faster spinning can lead to higher inertial stresses and subsequent device failure. 24 State - of - the - art flywheels have been engineered to operate for around 100,000 cycles. However, high costs ($8300/kWh) restrict flywheels to frequency regulation - related applications. 24 Figure 1 . 7 Market segments for energy storage devices in the electricity grid. Figure 1 . 8 Energy storage technologies employed for various stationary applications. 21 1.5 Market need for stationary sector The grid sector can be partitioned into three segments: generation, transmission and distribution and end - user ( Figure 1 . 7 ) and EES can be implemented in various levels s hown in Figure 1 . 8 . A 25 single EES technology is incapable of addressing the performance specifications and cost demands for all these segments in the grid sector. Henc e, a multitude of technology solutions address the varied needs and EES technologies primarily compete for mid - scale T&D and small - scale end - user applications. LIBs perform fairly well but are limited by high installation costs while the cheaper LABs do not offer good cycle - life characteristics. Thus, there is a market demand for a cost - effective EES technology in the stationary sector that can fill - in this gap between LIBs and LABs. The newly researched alternate battery chemistries are anticipated to a ddress this market need by having comparable performance to LIBs while still being relatively inexpensive. The Department of Energy (DOE) target specifications for EES devices are less than $150/kWh cost, at least 80% faradaic efficiency and minimum cycle - life of 5000, as shown in Table 1 . 1 . Current EES devices do not meet these specifications and the newly researched alternate chemistries are anticipated to enable dev elopment of low - cost, highly - efficient, long - life devices. Table 1 . 1 Short - term and Long - term DOE targets for EES technologies (stationary applications) . 25 Product Specification Short - term Target Long - term Target Cost <$250/kWh <$150/kWh Efficiency >75% >80% Cycle life 4000 5000 1.6 Need for Alternate Battery Chemistry Alternate battery chemistries, such as sodium and magnesium - based technologies, rely on earth - abundant natural resources. Cost - reductions are targeted by relying on widely available, inexpensive raw materials. It is expected that these low - cost devices wou ld help overcome the critical barrier for EES technologies into entering the various grid segments. Higher atomic weight 26 species like sodium and magnesium can be utilized for these applications since the grid sector does not place a large premium on attain ing high specific energy (Wh/kg). By 2100, the global lithium consumption by the transportation sector run in entirety by electric vehicles (with 90% recycling) is predicted to be 20 Mt. On the other hand, optimistic estimates for available lithium reso urce is 39 Mt. 26 In this best - case scenario, the demand for lithium mineral would be barely sufficient to meet the needs of automotive sector while omitting other potential applications like grid storage. 27 The relative abundance of sodium is significantly higher than that of lithium minerals as shown in Figure 1 . 9 . It is estimated that the available sodium resources in seen from Table 1 . 2 . Commercially, sodium is extracted from brine solutions (NaCl) using Solvay process, an important industrial method, to produce soda ash (Na 2 CO 3 ). Due to significance of soda ash to chemical industries (annual production of 11.6 MT), its monthly production rate is utilized as one of the economic indicators by the Federal Reserve Board. 28 Considering the enormity of the grid sector applications that scarcely utilizes batteries at the present, it is prudent to design a batte ry chemistry based on a cheap, mass - manufactured chemical commodity. 27 Figure 1 . 9 Relative abundance chart depicting plentiful sodium element as compared to lithium. 29 Table 1 . 2 Comparison of resources and physical properties of lithium, sodium and magnesium 30 Property Lithium Sodium Magnesium Crustal Abundance (ppm) 20 23,600 23,300 Oceanic Abundance (mg/L) 0.1 10,800 1,290 Resources (MT) 40 Unlimited Unlimited Redox Potential (V) vs. SHE - 3.04 - 2.71 - 2.37 Atomic Mass (amu ) 6.941 22.999 24.305 Atomic Radii ( pm ) 76 102 72 Ionic Charge 1+ 1+ 2+ Additionally, the majority of the lithium resources are restricted to few countries and geopolitics can potentially become an issue for ramping up production levels. However, sodium resources have a more uniform global distribution ( Table 1 . 2 ) and mineral costs would not be affected by geopolitical issues as much. Due to cheap and wide - scale mineral availability, sodium - based batteries offer the considerable potential for developi ng low - cost EES. This new EES technology 28 would help in product diversification of EES methods, potentially fill - in the market gap between LIBs and LABs and enable greater deployment of EES in various segments of the grid sector. 1.7 Rationale for investigating Na - based batteries 1.7.1 Na - ion vs Mg - ion batteries Sodium - and magnesium - based chemistries are highly researched as alternate battery chemistries for grid storage sector. 31 - 33 In terms of resource ava ilability, both sodium and magnesium are Table 1 . 2 . The standard reduction potential of sodium is closer to lithium (300 m V) than magnesium (600 mV). The more electropositive character of sodium than magnesium would permit building higher voltage cells with the former. Higher cell voltage translates to fewer cells/pack to achieve the product specification stack - level voltage and subsequently reduces failure modes owing to individual cell variability. Having fewer cells to monitor can also reduce the complexity in control systems at the pack level. Additionally, the divalent magnesium ions is known to suffer from extremely slo w solid - state diffusion due to strong interactions between magnesium ions and the host anions (larger coulombic attractive forces). 33 Hence, it is desirable to use monovalent intercalant species such as sodium to reduce the polarizing interactions and obtain reasonable fast solid - state diffusivity, essential for practical applications. The room temperature and intercalation - based operat ion of sodium - ion chemistries would eliminate the potential safety issues that are usually associated with high - temperature Na/S batteries. 29 1.7.2 Relative ease of scalability The high - volume manufacturing practices of building large lithium battery packs (~50 k Wh) are currently being optimized and this scaling effect is expected to produce marked cost - savings for transportation sector. A conventional battery manufacturing process is shown in Figure 1 . 10 . Figure 1 . 10 Manufacturing steps for advanced batteries such as Li - ion chemistry . Existing LIB plants can be modified to manufacture sodium - based batteries without the ne ed for new equipment or facility upgrades. The dryroom conditions currently used for LIBs would be suitable for sodium battery manufacturing. This transition can be performed by incorporating few key changes to the existing process modules as follows: i. Repl ace lithium - based active powders with sodium - module to fabricate sodium intercalation electrodes ii. Optimize electrolyte slurry viscosity and film thickness to specification levels in the de films of desired quality iii. Replace lithium - based salt with sodium - relatively fast sodium ionic conduction in the non - aqueous electrolytes 30 iv. Change organic solvent (linear and cyclic carbonate) type and to suit the requirements for sodium batteries By this relatively straightforward approach, the key cost saving measures of LIBs can be readily transferred to sodium - chemistries. This would also advance technology development from lab - scale research to building device prototypes to mass manufacturing in a fairly rapid fashion. 1.8 Insertion - based Batteries: Working Principle Insertion/Intercalation - based batteries function on the fundamental principle of a galvanic cell; electric current generated by two spontaneous, spatially separated redox reactions is utilized for doing external work. The redox reactions in insertion batte ries are facilitated by transition metal species that occupy specific crystallographic sites in an inorganic material. So, when the oxidation state of the transition metal changes, the concentration of the intercalant species gets modified by electrochemic al ion exchange process to maintain overall c harge neutrality . Thus, when using intercalation electrodes are utilized as anode and cathode, the intercalant species gets shuttled between them through the electrolyte media that allows for only ion transport . - Figure 1 . 11 using lithium/sodium as the intercalant species, and LiCoO 2 /NaCoO 2 as electrode materials. Insertion/r emoval of intercalant species in the host electrode framework can lead to changes in physicochemical and structural properties and even phase transformations. For attaining long cycle - life, the intercalant species needs to be shuttled between the two elect rodes for at least a few 1000 cycles while accommodating concomitant lattice stresses. Intercalation - based batteries that use solid - solution mechanisms do have electrodes undergo phase changes and typically offer excellent cycle life compared to other batt ery mechanisms (such as alloying). This is because the 31 host electrode materials retain their original structure fairly well and hence their intercalation capacity remains reasonably stable. Cycle life is a key performance specification for stationary appli cations and this research is focused only on intercalation - based materials. Figure 1 . 11 Working principle of an inserti on/intercalation - based battery. 34 1.9 Requisites for Intercalation - based Electrodes Designing high - performance, intercalation - type battery materials requires selecting electrodes with the following essential properties: i. High electr onic conductivity: Intercalation reactions involve exchange of intercalant species and a concomitant redox re action. Materials with high electronic conductivity would facilitate faster electron/hole hopping to the electrode - electrolyte interface where electrochemical reactions occur. 32 ii. High ionic conductivity: Intercalant species needs to diffuse fairly rapidly within the host material so that they can reach the electrode - electrolyte interface. This requires materials with inter - connected channels/tunnels where the intercalant species can reside. iii. Fast interfaci al kinetics: The interfacial reaction that occurs at the electrode/electrolyte interface allows needs to have fast heterogeneous reaction kinetics. iv. High redox potential: The redox potential of the electrode is directly proportional to the free energy of the formation between product (de - intercalated state for anode) and reactant (intercalated state for anode) as given by the Nernst equation (under electrochemical equilibrium) : the Pdt Rct v. Structural stability: Preserving the structure of the host materials during intercalation/de - intercalation is essential to retain the intercalant crystallographic sites. Significant changes in structure (phase transitions) can lead to loss of available intercalation sites and consequently poor charge storage capability. For instance, a material transforming from crystalline to amorphous would completely lose charge - storage capability due to intercalation mechanisms. vi. Stable Solid Electrolyte Interface (SEI): Insertion - electrode s typically operate in highly oxidizing and reducing conditions. Current Li - ion battery electrolytes rely on kinetic barriers to limit the thermodynamically favored reduction reactions at the anode (graphite). Formation of a stable, compact, electrically i nsulating and ionically conducting SEI as a passivation layer is essential to limit the side - reactions and attain stable capacity retention. 35 Finding suitable 33 electrolyte chemistries that can support different electrode materials by forming a stable SEI is one of the key challenges for realizing practica l Na - ion chemistry devices. 1.10 Research Problem Driven by the potential for stationary applications, Na - ion batteries has become a highly researched area in the recent years as highlighted by the rapidly growing number of research publications as shown in Figure 1 . 12 . 31 , 32 , 36 - 38 Identifica tion and characterization of new, high - performance functional materials that can support reversible sodium intercalation/de - intercalation reactions remains on e of the key challenges in Na - ion batter y field . Exploratory research of new inorganic host materi als forms the cornerstone for advanced batteries and can possibly open up new avenues for different application areas. This thesis research embarks on an exploratory research to identify potentially new inorganic host electrodes and characterize promising electrode candidates. The research goals include shortlisting promising compositions and increasing the understanding of material and electrochemical properties that would eventually lead to device prototyping and product commercialization. 34 Figure 1 . 12 Growing number of research publications on anode materials for Na - ion battery. 36 1.11 Thesis organization The second chapter of this dissertation begins with evaluating sodium metal SEI in different electrolytes and passivation study of current collectors. Thereafter, it discusses about the exploratory study of novel tunnel - type electrode materials: sodium nickel phosphate (Na 4 Ni 7 [PO 4 ] 6 ) and sodium cobalt titanate (Na 0.8 Co 0.4 Ti 1.6 O 4 ). The synthetic st rategies and electrochemical characterization of these materials are elaborated. The third chapter elaborates the characterization of sodium nickel titanate (P2 - Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 ) with prismatic co - ordination of sodium atoms. We evaluate the electrochem istry of this novel layer - structured material by selectively activating nickel and titanium redox couples. The bottleneck physicochemical processes are evaluated using impedance spectroscopy and the role of manganese substitution (P2 - Na 2/3 [Ni 1/3 Mn 1/3 Ti 1/3 ] O 2 ) on the electrical, ionic and electrochemical properties of these compositions are studied. 35 The fourth chapter discusses the structural characterization of sodium nickel titanate (P2) using experimental (neutron diffraction) and computational methods ( atomistic simulations). Rietveld refinement provides average structural information while force field models simulate average and local structural details. Computational methods were utilized to develop atomic density maps and the simulated structural patt erns were found to be consistent with the experimental results. The fifth chapter covers synthesis, structural and electrochemical characterization of sodium nickel titanate material (O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 ) with octahedral co - ordination of sodium atoms . We investigate the stabilizing influence of titanium atoms to mitigate poor capacity retention of O3 - type electrode materials and perform cycling - coupled - impedance studies, step - potential electroanalytical testing, and ex - situ XRD characterization. 36 BIBLIOGRAPHY 37 BIBLIOGRAPHY 1. Alley, R. et al. Basis. Cambridge Univ. Press 2. Tans, P. & Keeling, R. http://www.esrl.noaa.gov/gmd/ccgg/trends/. (2014). 3. Hansen, J., Ruedy, R., Sato, M. & Lo, K. Global Surface Temperature Change. Rev. Geophys. 48, 1 29 (2010). 4. Tett, S. F. B., Stott, P. A., Allen, M. R., Ingram, W. J. & Mitchell, J. F. B. Causes of twentieth - Nature 399, 569 572 (1999). 5. Liverman, D. From Uncertain to Unequivocal. Environ. Sci. Policy Sustain. Dev. 49, 28 32 (2007). 6. Dessler, A. & Parson, E. The sci ence and politics of global climate change . (Cambridge University Press, 2010). 7. Thomas, C. D. et al. Extinction risk from climate change. Nature 427, 145 8 (2004). 8. Araújo, M. B. & Rahbek, C. Ecology. How does climate change affect biodiversity? Scien ce (80 - . ). 313, 1396 7 (2006). 9. Thuiller, W. Climate change and the ecologist. Nature 448, 1 3 (2007). 10. World Econ. 7, 1 10 (2006). 11. Tol, R. S. J. The Economic Effects of Climate Change. J. Econ. Perspect. 23, 29 51 (2009). 12. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990 - 2013, Executive Summary . Annu. Rep. 1 27 (2013). 13. International Energy Statistics. http://www.eia.gov/cfapps/ipdbproject/iedindex3.cfm?tid=44&pid=44&aid=2& cid=ww, &syid=1980&eyid=2012&unit=QBTU (2015). 14. Beaudin, M., Zareipour, H., Schellenberglabe, A. & Rosehart, W. Energy storage for mitigating the variability of renewable electricity sources: An updated review. Energy Sustain. Dev. 14, 302 314 (2010). 15 . Denholm, P., Ela, E., Kirby, B. & Milligan, M. The Role of energy storage with renewable electricity generation . (2010). 38 16. Barnhart, C. J. & Benson, S. M. On the importance of reducing the energetic and material demands of electrical energy storage. En ergy Environ. Sci. 6, 1083 (2013). 17. REN21, Global Status Report 2014 (Renewable Energy Policy Network for the 21st Century); available at www.ren21.net/Portals/0/documents/e - paper/GSR2014/index.html. 18. Teleke, S., Baran, M. E., Bhattacharya, S. & Huan g, A. Q. Rule - Based Control of Battery Energy Storage for Dispatching Intermittent Renewable Sources. IEEE Trans. Sustain. Energy 1, 117 124 (2010). 19. Nature 496, 156 158 (2013). 20. Hadjipaschalis, I., Poullikkas , A. & Efthimiou, V. Overview of current and future energy storage technologies for electric power applications. Renew. Sustain. Energy Rev. 13, 1513 1522 (2009). 21. Electricity Energy Storage Technology Options; EPRI White Paper . (2010). 22. http://energ ystorage.org/energy - storage/technologies/flywheels. 23. Smart, M. C. & Ratnakumar, B. V. Effects of Electrolyte Composition on Lithium Plating in Lithium - Ion Cells. J. Electrochem. Soc. 158, A379 A389 (2011). 24. Tzeng, J., Emerson, R. & Moy, P. Composite flywheels for energy storage. Compos. Sci. Technol. 66, 2520 2527 (2006). 25. Grid Energy Storage: Executive summary; U.S. DOE . (2013). 26. Gruber, P. W. et al. Global Lithium Availability. J. Ind. Ecol. 15, 760 775 (2011). 27. Tarascon, J. - M. Is lithium t he new gold? Nat. Chem. 2, 510 (2010). 28. http://minerals.usgs.gov/minerals/pubs/commodity/soda_ash/. 29. Rare Earth Elements - Critical Resources for High Technology; USGS Fact Sheet 087 - 02 . (2002). 30. CRC Handbook of Physics and Chemistry, 77th Edition . 31. Slater, M. D., Kim, D., Lee, E. & Johnson, C. S. Sodium - Ion Batteries. Adv. Funct. Mater. 23, 947 958 (2013). 32. Palomares, V. et al. Na - ion batteries, recent advances and present challenges to become low cost energy storage systems. Energy Environ. Sci. 5, 5884 5901 (2012). 39 33. Levi, E., Gofer, Y. & Aurbach, D. On the Way to Rechargeable Mg Batteries: The Challenge of New Cathode Materials. Chem. Mater. 22, 860 868 (2010). 34. http://neutronsources.org/news/scientific - highlights/towards - sodium - ion - ba tteries - understanding - sodium - dynamics - on - a - microscopic - level.html. 35. Peled, E., Golodnitsky, D. & Penciner, J. Handbook of Battery Materials . (2011). 36. Dahbi, M., Yabuuchi, N., Kubota, K., Tokiwa, K. & Komaba, S. Negative electrodes for Na - ion batterie s. Phys. Chem. Chem. Phys. 16, 15007 (2014). 37. Pan, H., Hu, Y. - S. & Chen, L. Room - Temperature Stationary Sodium - Ion Batteries for Large - Scale Electric Energy Storage. Energy Environ. Sci. 6, (2013). 38. Ellis, B. L. & Nazar, L. F. Sodium and sodium - ion e nergy storage batteries. Curr. Opin. Solid State Mater. Sci. 16, 168 177 (2012). 40 2. S YNTHESIS, AND ELECTROCHEMICAL CHARACTERIZATION OF NOVEL TUNNEL - TYPE SODIUM INTERCALATION M ATERIALS 2.1 Introduction Tunnel - type inorganic battery materials can be classified into two structural categories: polyanion - and spinel - based. Polyanion - based electrode materials have tetrahedrally co - ordinated polyanionic functional groups (XO 4 ) n - Some of the well - studied polyanionic materials include LiMPO 4 (M - Fe, Mn, Ni, Co), Li 2 MSiO 4 (M - Fe, Mn), LiMSO 4 F (M - Co, Ni), etc. 1 - 6 These materials possess covale ntly bonded oxygen atoms that provide excellent charging stability (especially at high potentials) and improved thermal stability. 1 , 7 Hence, these polyanionic materials are intrinsically safer, an important feature for large - scale grid storage applications. Additionally, suitable selection of polyanion functional groups allows for selective tuning of the redox potential of the transition metal species. For instance, the transition metal in NASICON - structured materials have shown a wide range of redox potential values solel y based on the type of polyanion framework. 8 - 10 The bonding nature of the polyanion groups affect the ionicity of the transition metal - oxygen bonds and hence alters the redox potential. For Na - ion batteries, polyanionic materials such as Na 2 MPO 4 F (M - Fe, Mn), Na 2 MP 2 O 7 (M - Fe, Co, Mn, Cu), NaFePO 4 , Na 3 M 2 (PO 4 ) 3 (M - V, Ti) , Na 3 M 2 (PO 4 ) 2 F 3 (M - Ti, Fe, V) have shown promise. 11 - 19 In this section, Sodium nic kel phosphate (SNP), with stoichiometry Na 4 Ni 7 (PO 4 ) 6, is investigated as a potential tunnel - type electrode to take advantage of the structural stability of these classes of materials. 41 Spinels, with general stoichiometry AB 2 O 4 , have been shown to possess a ttractive intercalation/de - intercalation properties using compositions such as LiMn 2 O 4 and Li 1 [Li 0.33 Ti 1.67 ] O4. 20 - 24 This piqued our interest in exploring similar prototypical structures with sodium as intercalant species with spinel - related host framework. However, spinel structures are not favorable for sodium - based quaternary oxides due to the large - size of the sodium atoms for the - based prototypical structures with single or double tunnels depending that allows for a longer Na - O bond. 25 Sodium cob alt titanate (SCT), with stoichiometry Na 0.8 Co 0.4 Ti 1.6 O 4 , has calcium ferrite - type structure and is investigated as an alternative tunnel - type electrode candidate. This chapter discusses about synthesis and electrochemistry of two tunnel - type materials wit h compositions Na 4 Ni 7 (PO 4 ) 6 (polyanionic - based) and Na 0.8 Co 0.4 Ti 1.6 O 4 (calcium ferrite - based). Both these tunnel - type materials contain electrochemically - inert units, in the form of phosphate tetrahedra and titanium octahedra, which interconnect to form a relatively stable host - framework with possibly improved structural stability. Sodium ion conduction would occur in the in 3 - D interconnected tunnels that lie within the host framework. 2.2 Structural features 2.2.1 Sodium nickel phosphate: Na 4 Ni 7 (PO 4 ) 6 Previously, the crystal structure of Na 4 Ni 7 (PO 4 ) 6 was refined using X - ray diffraction with single - crystals grown by flux method. 26 experimental diffraction dataset and nickel and phosphorus were assigned octahedral and tetrahedral co - interconnecte 42 - - turn linked by corner - shared phosphorus polyhedra resulting in parallel tunnels/cavities that hosts sodium atoms ( Figure 2 . 1 ). The presence of continuous tunnels and unoccupied sodium sites in this phosphate material provides the feasibility for reasonably facile solid - state sodium ion transport. We anticipate that t - electrochemical de - sodiation. Figure 2 . 1 Crystal structure o f Na 4 Ni 7 (PO 4 ) 6 ; Pink polyhedra represents nickel octahedra, green polyhedra represents phosphate tetrahedra and yellow spheres indicate sodium atoms. Oxygen atoms are not included for clarity. 26 2.2.2 Sodium cobalt titanate: Na 0.8 Co 0.4 Ti 1.6 O 4 Na 0.8 M 0.4 Ti 1.6 O 4 was reported to crystallize with space group Pnma. 25 The studied composition, Na 0.8 Co 0.4 Ti 1.6 O 2 , has edge - shared groups of four transition metal octahedra (blue polyhedra) - to the b - axis infinitely as shown in Figure 2 . 2 . It is likely that these quaternary oxides support facile sodium ion transport due to their low activation energy (1. 1 eV) for sodium hopping and unusually high sodium thermal vibrational parameter (U 22 ). 25 , 27 43 Figure 2 . 2 Crystal structure of Na 0.8 Co 0.4 Ti 1.6 O 4 ; blue polyhedra represents transition metal octahedra (cobalt and titanium) and yellow spheres represent sodium atoms; oxygen atoms are omitted for clarity. 25 2.3 Inorganic Synthesis The synthetic conditions for the quaternary oxide phases were optimized to yield powders of desirable morphology for fa brication of composite, porous electrode films. Battery electrodes typically consists of powders with relatively small - sized particles (preferably in the length - scale of 100s of micrometer to nanometers) to avoid potential transport limitations. High - tempe rature powder processing can lead to particle sintering and this should be avoided. 2.3.1 Sodium nickel phosphate synthesis Previous report of SNP synthesis was carried out by slow cooling of high temperature melts (1150 o C) to obtain single crystals for diffra ction studies. 26 We developed a solid state synthetic method to produce phase - pure powders at a much lower temperature. The precursors Na 2 CO 3 (S4132, Sigma - Aldrich, 99.5%), NiO (399523, Sigma - Aldrich, 99%) and NH 4 H 2 (PO 4 ) 2 (204005, Sigma - Aldrich, 99.999%) were dry milled in the target stoichiometry 44 using a high - energy SPEX mill (8000M Mixer/Mill, SPEXSamplePrep) with steel balls inside a hardened steel vial. The homogenized powders were fired with an alumina crucible (Coors TM high - alumina crucible, Sigma - Aldrich) inside a box furnace (CWF - 1300, Carbolite and Lindberg blue M, Thermo - scientific model s) to provide sufficient thermal energy for activating the reaction and enabling faster diffusion, as typical in solid - state reaction technique. The overall synthesis schematic is illustrated in Figure 2 . 3 . Figure 2 . 3 Schematic of solid state reaction technique. Powder X - ray diffraction experiments were performed using Bruker Da - vinci D8 Diffractometer with copper X - ray source at the Center for Materials Characterization, Department of Chemistry, Michigan State University. Crystalline, phase - pure powders were obtained by heating the powders to 900 o C for 10 h duration as revealed by the X - Ray Diffraction (XRD) patter ns that show sharp, intense peaks and a good match with the database standard (ICSD reference code: 84 - 0845) as seen in Figure 2 . 4 . 45 Figure 2 . 4 XRD characterization of sodium nickel phosphate prepared by solid state reaction. The powder morphology was analyzed using JEOL - 7500F ultra - high resolution SEM equipment at the Center for Advanced Microscopy (CAM), MSU. Cond uctive - adhesive backed carbon tape (Ted Pella, Inc.) is placed on top of an aluminum stub and inorganic powders are sprinkled on top. The powders are gently tapped - off to dislodge excess powder accumulation. Typically, the powders are coated with a thin, s puttered gold film (~10 - 20 nm) to mitigate charging issues that can occur due to the interaction between electron gun and sample powders. The 7500F SEM is operated under ultra - high vacuum mode at sample chamber pressure of 96 µ Pa. The SEM images showed tha t the as - synthesized powders consisted of large (>100 µm), dense particles that are not ideal for electrochemical characterization. The individual particles did not have any visible pores and appears sintered ( Figure 2 . 5 a and b). Attempts were made to prepare smaller - sized powders by wet - chemistry methods. 46 (a) (b) Figure 2 . 5 (a) Sodium nickel phosphate powder morphology prepared by solid state reaction, and (b) Higher - magnification image of a particle showing dense morphology. Sol - gel synthesis facilitates atomic - level mixing by incorporating all the metallic species in a gel network. 28 - 30 Having chemical homogeneity at an atomic - scale typically enables reduction of phase formation temperature and consequently helps to avoid undesired particle agglomeration and sintering. Water soluble precursors such as CH 3 COONa (S8750, Sigma - Aldrich, 99%), Ni(CH 3 COO) 2 .4H 2 O (72225, Sigma - Aldrich, 99%) and NH 4 H 2 (PO 4 ) 2 (204005, Sigma - Aldrich, 99.999%) were utilized while citric acid (36664, Alfa - Aesar, 99.5%) functioned as the chelating agent and nitric acid (438073, Sigma - Aldrich, 70%) served to maintain acidic pH. Evaporation of water res ulted in the formation of transparent green gels as shown in Figure 2 . 6 . The green gel was firing at 800 o C to burn the organic components (mostly from citric acid) and promote phase formation and this leads to the formation of yellow powders. 47 Figure 2 . 6 Schematic of sol - gel synthesis technique . Figure 2 . 7 XRD characterization of sol - gel synthesized powders of sodium nickel phosphate. XRD patterns confirm the formation of the target phase when firing duration is at least 4 hours (ICSD reference code: 84 - 0845) . It was found that reducing the synthesis duration to 2 hours results in an impurity phase formation (NaNiPO 4 , ICSD reference code: 34 - 1446), as shown in Figure 2 . 7 . The sol - gel powders had a smaller particle size (~15 µm) than the powders prepared 48 by solid - state reaction due to the lower firing temperature enabled by chemical homogeneity at atomic - scale. The SEM images , in Figure 2 . 8 , show relatively large primary particles comprising of partially sintered secondary particles. The latter were found to be ~ 1 µm in diameter and are interconnected to form an extensive pore network through which the liquid electrolyte can possibly perm eate and facilitate faster ion diffusion. (a) (b) Figure 2 . 8 (a) Sodium nickel phosphate powder morphology prepared by sol - gel reaction, and (b) Higher - magnification image of a particle showing porous morphology. 2.3.2 Sodium cobalt titanate synthesis SCT was synthesized by solid - state reaction method using NaNO 3 (S5506, Sigma - Aldrich, 99%), Co(CH 3 COO) 2 (208396, Sigma - Aldrich, 99.5%), and TiO 2 (248576, Sigma - Aldrich, 99%, Anatase phase) as starting precurso rs. The precursors were wet milled using high - energy SPEX mill with steel balls as grinding media and 2 - propanol as solvent. The well - mixed slurry was dried using an IR lamp in a glass petri dish and the dried powders were fired in a box furnace at 900 o C f or 10 hours. Na - Co - Ti - O composition with calcium - ferrite prototype structure is not reported in the ICSD database. XRD patterns, in Figure 2 . 9 a, of the synthesized po wders shows excellent 49 match with Na 0.79 [Fe 0.8 Ti 1.6 ]O 4 (ICSD reference code: 87 - 1550), that is isostructural with calcium - ferrite double tunnel structure. Since replacing cobalt in place of iron did not change the space group Pnma, we propose that the prepa red powders retain the tunnel structure. The as - synthesized powders were found to be ~3 µm sized rods as primary particles that appeared non - porous. The particles do not show any evidence of sintering ( Figure 2 . 9 b) while maintaining good crystallinity suggesting that the synthetic conditions are fairly close to optimal conditions. Hence, no attempts were made to prepare the powders by sol - gel method since the researc h objective was to have a fairly straightforward synthetic protocol (to enable scalability and cost effectiveness). (a) (b) Figure 2 . 9 (a) XRD characterization of sodium cobalt titanate prepared by solid state reaction, and (b) SEM morphology of as - synthesized sodium cobalt titanate powders. 50 2.4 Film fabrication 2.4.1 Substrate selection Highly - conductive metal substrates (or current collectors), typically aluminum (MTI Corp., EQ - bcaf - 15u - 280) and copper (MT I Corp., EQ - bccf - 25u), are utilized for backing the active components of the electrode films. In commercial devices, these substrates carry the current from individual electrodes to the cell terminals. The prerequisite properties for substrates include goo d electrochemical stability (oxidizing environment for cathodes and reducing conditions for anodes), suitable chemical stability with active components (oxide materials, lithium metal, sodium metal, etc.), excellent electrical conductivity (to avoid ohmic losses), good thermal conductivity (for heat dissipation) and sufficient mechanical properties (to withstand thermal vibrations, mechanical impacts typical during battery operation). Aluminum substrate is utilized for casting the cathode films as it provides excellent electrical conductivity and remains stable in highly oxidizing cathodic conditions (due to passivation films). However, phase diagrams suggest that aluminum forms binary alloys such as AlLi, Al 2 Li 3 , Al 4 Li 9 31 with lithium metal and hence i s not a suitable substrate for lithium battery anodes. Aluminum substrates can be used for Na - ion batteries as it does not alloy with sodium metal. The low - cost and light - weight aluminum substrates are expected to provide energy density gains and cost redu ction opportunities at pack - level battery configurations. 2.4.2 Casting methods The as - synthesized powders are cast into a porous, composite electrode for electrochemical characterization. The film electrode is comprised of the active powders (quaternary oxides) , conductive carbon (Super C65, Timcal, USA), and polyvinylidene fluoride (44080, Alfa - Aesar) in 51 80:10:10 weight ratio. The active powders and conductive carbon are dispersed thoroughly in a viscous solution made from dissolving polyvinylidene fluoride in N - Methyl - 2 - pyrrolidone (L03595, Alfa - Aesar) solvent using high - energy ball milling. The wet slurry is casted onto the aluminum substrate using an adjustable film applicator (doctor - blade, EQ - Se - KTQ - 100, MTI Corp.) and film coater (MSK - AFA - III, MTI Corp.) a fter selecting the desired film thickness and setting the traverse speed at 10 mm/s. Suction is created by an oil - free vacuum pump (Model no. - Carr) to the aluminum vacuum chuck and this secures the aluminum foil/substrate in place. A schematic of film fabrication is illustrated in Figure 2 . 10 . The film is dried using an IR lamp overnight to ensure that the using a hollow - punch (Mayhew tools, 66000) and transferred to MBraun labmaster - SP glovebox (with <1 ppm O 2 and <1 ppm H 2 O) to minimize exposure to moisture and ca rbon dioxide. Figure 2 . 10 Schematic of doctor - blade slurry casting method for fabricating battery electrodes. 52 2.5 Electrochemical characterization 2.5.1 Cell setup and testing Electrochemical testing is performed in 3 - electrode configuration which includes working electrode, counter electrode and a reference electrode. Working electrode consists of an electrode nt terminals (equivalent to plate terminals in commercial packs). The counter and reference electrodes are prepared by cutting soft sodium ingots into thin slices with a knife, punching the slices into a circular disks and pressing the disks against a copp er rod with Teflon tape and manual press. The three electrodes are inserted into the three legs of a Swagelok tube fitting (PFA - grade, Swagelok Inc.,) while inserting a polyethylene separator between the rods to avoid cell shorting. Figure 2 . 11 depicts electrode arrangement in cell assembly. The cell assembly is performed inside glovebox and Swagelok assembly ensures hermetic sealing of cell components. Cell testing is carried out using Biologic Potentiostat VSP300 and SP200 models outside the glovebox. The accuracy of the potentiostats was periodically tested using standard dummy cells from Biologic. Figure 2 . 11 Schema tic of Swagelok 3 - electrode cell setup for electrochemical testing. 53 2.5.2 Evaluation of the electrolyte and current collectors Sodium metal is frequently used in lab - scale half - cell tests due to the unavailability of commercial high - performance sodium - based ano des as counter electrodes. We tested the electrochemistry of Na 2 Ti 3 O 7 , a reported low - voltage anode (E o =0.3 V vs. Na/Na + for Ti 4+/3+ ) for Na - ion batteries. 32 However, the performance of the electrodes fabricated using the commercial powders (401307, Sigma - Aldrich) were found to be unsatisfactory with reversible capacity of only ~10% of theoretical. XRD testing revealed the commercial powders to contain a significant amount of Na 2 Ti 6 O 13 impurity phase in addition to the target Na 2 Ti 3 O 7 phase. Hence, further attempts were focused on finding a suitable electrolyte combination that can potentially work with the reactive sodium metal for the scope of preliminary electrochemical testing. The stability of sodium metal (282065, Sigma - Aldrich) in two electrolytes 1M NaClO 4 in EC/DMC (50/50 vol%) and 0.5M NaPF 6 in EC/DEC (50/50 vol%) were evaluated using impedance spectroscopy. The dry sodium metal sticks packed in nitrogen atmosphere is opened and stored inside the glovebox atmosphere. During ce ll assembly the outer regions of the stick are removed to discard the white coating (possibly oxide) and cut into ~3 - 5 mm slices using a steel knife. The thin slides are placed between two separator pieces and manually pressed (with Dayton ½ ton Arbor pres s). The pressed - slices are punched into disks with a hollow punch manually. Symmetric cells were assembled using two sodium metals that were affixed to copper metal rods in 2 - electrode configuration. Impedance tests were carried out by applying a sinusoidal voltage perturbation of 10 mV amplitude over the open - circuit potential in a frequency range of 1 and the space between the rods was filled - in with surplus electrolyte solution. The 1 st impedance 54 test (black curve) was performed after 4 hours of open - circuit condition and subsequent tests every 2 hours. Impedance spectrum reflects the evolution of the Solid Electrolyte Interface (SEI) on sodium metal surfac e over time without any significant polarization from open circuit conditions. Similar analysis has been reported for lithium metal SEI formation. 33 Perchlorate - based electrolyte exhibits a single semi - arc in t he initial cycles that can be attributed to passivation films formed due to the degradation of organic solvents as shown in Figure 2 . 12 a. (a) (b) Figure 2 . 12 Impedance spectroscopy curves for sodium metal SEI formation in (a) 1M NaClO4 in ethylene carbonate/dimethyl carbonate, and (b) 0.5M NaPF6 in ethylene carbonate/diethyl carbonate; 1st cycle refers to test aft er 4 hours open - circuit and subsequent tests. Prolonged exposure results in the formation of a 2 nd arc related to the formation of a secondary passivation film. The increasing film resistance is due to continuous incorporation of the electrolyte degradation components and subsequent increase in film thickness. Fluorine - based 55 electrolyte shows a smaller single - arc over the tested time - duration and the film resistance increase is more gradual as seen in Figure 2 . 12 b. We attribute this effect to the formation of a more stable passivation film involving fluorine - related species such as sodium fluoride as previously reported with a fluorinated solvent. 34 The electrolyte resistance did not change significantly under these test conditions as the cell - configuration was flooded and the byproducts did not acc umulate to a significant extent. These impedance spectra were obtained under steady - state open circuit conditions to investigate the reactivity of sodium metal under open circuit conditions. The nature of the passivation films would change during actual c ell operation that includes current perturbation. Also, during galvanostatic testing, sodium metal plating is visually observed to be in the form of dendritic powders. These dendrites lose eventually contact with sodium metal due to poor adhesion and resul ts in loss of partial deactivation (at the counter electrode). During long - time cycling tests, these dendrites also potentially pose concerns of cell shorting through the separator. Significant electrolyte optimizations are further required to address thes e issues and further technology commercialization. Current collectors and solvent species can typically get oxidized during the electrochemical testing of the inorganic materials and this can result in contamination of the electrolyte. Electrolyte contami nation can lead to additional redox contributions, complex interactions with studied active components and should be avoided. Electroanalytical techniques such as cyclic voltammetry are routinely employed to define the operable voltage window. This techni que involves scanning over the desired voltage window at a predefined scan - rate (mV/s) and measuring the current response. A distinct current increase would 56 be noticeable corresponding to the electrochemical reactions of solvent species and current collect ors. Using multiple scans, we can investigate the suppression capability (corrosion resistance) of the surface passivation films that typically form in the first cycle for various materials. Different substrates such as aluminum, anodized aluminum, stainle ss steel (grade 316), titanium and graphite were evaluated for their electroanalytical response as shown in Figure 2 . 13 and Figure 2 . 14 around 3 - 4 inches long. The anodized aluminum (hard anodized 6061 Al rod, McMaster - Carr) has a black insulating oxid e coating on the rod circumference while the end cut - sections have exposed aluminum metal (to provide electrical contact to electrode films). The oxide coating has excellent corrosion resistance and is left intact to minimize the electrolyte contact with b are aluminum surface. The rods were polished with 600 grit sanding sheets (polished finish grade, McMaster - Carr) to obtain a fresh metal surface with smooth finish while removing surface impurities. Subsequently, they were subjected to ultrasonication (usi ng Branson, 3510 sonicator) in 2 - propanol solution for ~15 minutes to dislodge any residual metal powders from prior sanding step and air - dried before cell - assembly. The 1 st cycle cyclic voltammogram of aluminum exhibits a peak current of 54 µ A during th e oxidative scan and this gets reduced substantially (10 th cycle peak current ~ 6 µ A) in the subsequent cycles due to surface passivation that limits further reactions. Anodized aluminum rod shows much lower peak current (9.9 µ A) in the 1 st cycle due to lo wer exposed metal surface contacting the electrolyte. It also has a strong passivation effect during subsequent cycles (10 th cycle peak current - 0.7 µ A). 57 (a) (b) (c) (d) Figure 2 . 13 Cyclic voltammograms scanned between 1.5 - 4.2 V voltage window for (a) aluminum, (b) andoized aluminum, (c) stainless steel 316, and (d) titanium. 58 (a) (b) Figure 2 . 14 (a) Cyclic voltammogram of graphite scanned between 1.5 - 4.2 V voltage window, and (b) oxidative voltammogram peak currents for 1 st and 10 th cycle. Stainless steel (grade 316), titanium and graphite have 1 st cycle peak current values of 140 µ A, 64 µ A, an d 106 µ A respectively. Like aluminum, these materials also show a tendency towards passivation as evidenced by peak current reduction as shown in Figure 2 . 14 b . For further testing, only anodized aluminum is utilized as current collector as it exhibited superior oxidative stability compared to the remaining materials. We also tested the anodized aluminum rod under galvanostatic conditions that are typically us ed in battery electrode testing. The galvanostatic results are in accord with cyclic voltammetric studies. At 23.6 µ A/cm 2 current density, the upper voltage limit of 4.5 V is reached in 6 min during which aluminum passivation occurs. Subsequently, voltage profiles are steeper as further oxidation reactions are not supported readily and the rods can be considered to be fairly inert under the test conditions ( Figure 2 . 15 ). 59 Figure 2 . 15 Voltage profile of anodized aluminum current collector tested under galvanostatic condition with 4.5V cutoff using 3 - electrode setup. 2.5.3 Validation of electrochemical methods using standard intercalation materials : Na 0.7 CoO 2 and LiFePO 4 In order to validate the experimental methods for film fabrication and cell assembly, well - studied materials such as LiFePO 4 and Na 0.7 CoO 2 were tested. The goal was to ensure that the experimental results are consistent with the literature reports. LiFePO 4 powders were synthesized by dry milling Li 2 CO 3 , NH 4 H 2 PO 4 and Fe(C 2 O 4 ).2H 2 O (from Sigma - Aldrich) with 10 wt% carbon mixture (equal proportions of sucrose and Ketjen Black, AkzoNobel) and firing the precursors in a tube furnace with argon atmosphere at 600 o C for 6 hours duration. This modified procedure was adopted from literature results suggesting that the incorporation of carbon additives improves electrochemical performance. 35 The as - synthes ized powders were comprised of <100 nm sized particles that are interconnected by a carbon network. Na 0.7 CoO 2 was synthesized by straightforward solid - state reaction method in a box furnace. Materials were subject to electrochemical testing using 3 - electro de cell assembly. 60 Figure 2 . 16 a shows the charge/discharge curves for LiFePO 4 electrodes using lithium metal as counter and reference electrodes and LiPF 6 - based comme rcial electrolyte from MTI Corp. (EQ - LBC3051C). A flat - voltage profile at 3.4 V vs. Li/Li + is observed demonstrating the activity of Fe 2+/3+ redox couple accompanied by a two phase reaction (involving LiFePO 4 and FePO 4 phases) matching the literature repor ts. 2 , 35 , 36 The discharge capacity values are 74% of the theoretical capacity (170 mAh/g) at C/10 rate. Na 0.7 CoO 2 electrodes were tested with sodium metal as counter and reference electrodes and 0.5 M NaPF 6 in EC/DMC electrolyte (50/50 vol%). The complex voltage steps and plateau, in Figure 2 . 16 b are attributable to single/bi - phasic domains and vacancy ordering as previously reported. 37 These results provide a good confidence on our experimental design as they replicate the literature reports fairly accurately. (a) (b) Figure 2 . 16 Voltage profiles of (a) lithium iron phosphate under galvanostatic conditions at C/10 rate demonstrating two - phase reaction between FePO 4 and LiFePO 4 , and (b) Na 0.7 CoO2 showing complex voltage steps and plateaus due to multiple phase transitions. 61 2.5.4 SNP Electrochem istry: Galvanostatic testing SNP has a theoretical capacity of 106 mAh/g for complete sodium de - intercal ation (x=4) from the structure [ Na 4 Ni 7 ( PO 4 ) 6 to Ni 7 ( PO 4 ) 6 ] . The sodium de - intercalation/intercalation reactions supported by Ni 2+/4+ re dox couple can be represented in Equation 2 . 1 . 2 . 1 Galvanostatic charge/discharge studies are executed to get an improved understanding of the reversibility of the redox reactions. Studied electrolytes include 1M NaClO 4 in Ethylene Carbonate/Dimethyl Carbonate (EC/DMC) and 0.5 M NaPF 6 in Ethylene Carbonate /Diethyl Carbonate (EC/DEC) solutions in voltage window between 3.2 and 4.2 V. During charging, there is an initial rapid voltage rise to 2.8 V vs. Na/Na + after which the voltage curve shows a gradual sloping profile, as depicted in Figure 2 . 17 a. Prolonged charging results in continued extraction of sodium ions resulting in concomitant generation of additional Ni 4+ states. Complete sodium de - intercalation could be achi eved after which the voltage starts to rise due to electrolyte oxidation reactions to support the current. The current direction was reversed to induce reduction of Ni 4+ species but this resulted in a significant voltage polarization with very little sodi um intercalation with voltage cutoff limited to 1.5 V. Lowering the voltage cutoff to 0.8 V leads to a long voltage plateau which cannot be accounted for by just considering sodium intercalation reactions. It is likely that in addition to sodium intercalat ion, there is considerable amount of electrolyte reduction reactions, possibly catalyzed by SNP particles. Nevertheless, the subsequent charging cycle provides ~50% theoretical capacity at a lower electrode potential (~2.0 V vs. Na/Na + ) in Figure 2 . 17 b. The voltage profile remains in the same regime for further cycles and the difference between the first and second cycle can possibly be accounted for by local re - 62 organiz ation effects (not captured by XRD discussed in the next section). Future studies to increase the understanding of voltage hysteresis effect (1 V) and electrocatalytic reduction of organic solvents are required to improve the energy density and arrest capa city fade. (a) (b) Figure 2 . 17 Sodium nickel phosphate tested under galvanostatic conditions with discharge cut - off voltage of (a) 1.5 V in 1st cycle, and (b) 0.8 V. 2.5.5 SNP Ex - situ XRD characterization Ex - situ XRD testing is utilized to characterize the film electrodes at various sodium contents. This helps in detecting electrochemically - driven phase transitions and analyzing the crystallinity of the studied battery materials. The film e lectrodes are subjected allowed to reach different charge/discharge voltages under galvanostatic conditions to achieve various sodium intercalation contents inside the electrochemical setup. After passing charge and allowing for equilibration (typically 2 hours), the cell is disassembled to obtain the electrodes and the film is washed with DMC solvent to remove any residual salt components inside the glovebox. The washed film is dried under ambient conditions and transferred to a vial for XRD characterizati on. The film is mounted on Bruker sample holder and secured using wax. Mylar tape is placed over the sample 63 holder and a seal mechanism limits air exposure during the characterization. Inclusion of wax and Mylar does lead to unwanted bumps and peaks in the XRD patterns ( Figure 2 . 18 ) and there were found not to overlap with the main peaks of the target SNP phase. Figure 2 . 18 Ex - situ XRD characterization of SNP electrode film at different state - of - charge, Mylar film and wax (used for sticking the electrode to the holder base). We found that the 3.2 V charged electrode peaks are identical to those of pristine powders de monstrating that the SNP material retains the same crystal structure even after complete sodium de - intercalation, as shown in Figure 2 . 18 . We attribute this structural effect of phosphate functional groups, typical of polyanionic materials. The electrodes subjected to deep discharge of 1.2 and 1.0 V also have comparable patterns indicating that the redox reactions occurring at those potentia l regions are not due to the formation of metallic nickel species (conversion - type mechanisms). We demonstrate that the reaction mechanism that provides charge storage capability is due to the desired intercalation - type mechanisms. Since SNP has a relative ly 64 low redox potential (~2 V) and only a modest reversible capacity (~ 50 mAh/g), we focused our attempts to investigate different structural types for our exploratory research. 2.5.6 SCT Electrochemistry: Potentiostatic Intermittent Titration Testing (PITT) PITT involves scanning the electrode potential over the working range (upper and lower cutoff voltages) with voltage steps and allowing for current relaxation (C/50) during each step. The relaxation step allows the material to approach close - to equilibrium conditions. For complete sodium de - intercalation (x=0.8), the theoretical capacity is calculated to be 117 mAh/g. Sodium intercalation/de - intercalation reactions due to Co 2+/4+ r edox couple can be represented in Equation 2 . 2 . 2 . 2 SCT was tested using 0.5 M NaPF 6 in 50/50 vol% EC/DEC electrolyte in a 3 - electrode cell assembly. The 1 st charging voltage profile slopes gradually between 3.6 and 4.2 V and remains fairly identical for next two cycles as shown in Figure 2 . 19 a. The discharge profile also follows a sloping profile with a large initial voltage polarization. The symmetric nature of the charging and discharging curves suggests reversible sodium intercalation/de - intercala tion reactions. The capacity retention was also found to be fairly stable in the tested 3 cycles and the faradaic efficiency improves to 86% in the 3 rd cycle ( Figure 2 . 19 b). However, there is a large voltage hysteresis of 1.4 V between charge and discharge curves even while testing the cell at 70 o C. 65 (a) (b) Figure 2 . 19 (a) Voltage curve of sodium cobalt titanate f rom PITT with C/50 current limit, 100 mV step potential tested at 70 o C using 0.5 M NaPF6 in EC/DEC electrolyte, and (b) Discharge capacity and faradaic efficiency with cycle number. SCT was also evaluated with 1M NaClO 4 in EC/DMC electrolyte using slow - rate galvanostatic testing of C/100 at room temperature. The voltage hysteresis increases to 1.6 V for room - temperature testing suggesting a thermally activated limiting process ( Figure 2 . 20 a). We also show that the hysteresis is not related to the nature of solvent and is inherent to the studied material. In Figure 2 . 20 b, Ex - situ XRD results confirm that the SCT pristine peak positions are fairly comparable to those of partial charged (20%) and discharged films. The absence of additional peaks indicates that there is not a tendency towards phase transformations. The partially de - i ntercalated electrode has slightly different peak intensities and suggesting the possibility of sodium redistribution amongst the remaining vacant sites. 66 (a) (b) Figure 2 . 20 (a) Voltage profile of sod ium cobalt titanate tested under at C/100 rate using 1M NaClO4 in EC/DMC electrolyte at room temperature, and (b) Ex - situ XRD of SCT at pristine, 20% charged and discharged state. We hypothesize that the hysteresis phenomenon is related to relatively slow diffusion of sodium ions in the tunnels and further improvements can be realized by grain - size reduction (nano - scale) by chimie douce synthetic methods. The scope of this exploratory research was to identify potential inorganic materials that can poss ibly sustain relatively facile sodium intercalation/de - intercalation reactions using micron length - scale particles. The reported tunnel - type materials does show promise in terms of structural stability (after sodium de - intercalation) but their electrochemi stry is fairly constrained by diffusional limitations. We initiated the study of novel layer - structured, bi - metallic transition metal oxides, with the general composition Na x M y 1 - y O 2 as suitable sodium intercalation electrodes in the subsequent chapters. 2.6 Summary In this chapter, two tunnel - type materials Na 4 Ni 7 (PO 4 ) 6 and Na 0.8 Co 0.4 Ti 1.6 O 2 were explored as potential sodium intercalation electrode materials. These framework materials were anticipated to 67 possess good stability based on structural connectivi ty and reasonable sodium transport through the 3 - D tunnels. Phase - pure, crystalline powders of these two compositions were successfully synthesized by solid - state reaction and sol - gel techniques as demonstrated by the characteristic XRD peak sets. SNP show s a sloping charging voltage profile at 2 V and discharge profile at 1.25 V with a noticeable voltage hysteresis. These reactions were confirmed to be due to topotactic exchange of sodium intercalant without any phase change. SCT was also found to be elect rochemically active with charging and discharging voltage profile at 3.7 and 2.5 V, respectively when tested at 70 o C. The large voltage hysteresis is predicted to be due to slow - diffusion of sodium. These materials show promise in terms of structural stabi lity as no phase change was evident upon sodium removal. However, to mitigate voltage hysteresis, future research needs to focus on developing improved chimie douce methods for nano - sizing the particles without particle sintering. 68 BIBLIOGR APHY 69 BIBLIOGRAPHY 1. Gong, Z. & Yang, Y. Recent advances in the research of polyanion - type cathode materials for Li - ion batteries. Energy Environ. Sci. 4, 3223 (2011). 2. Padhi, A. K., Nanjundaswamy, K. S. & Goodenough, J. B. Phospho - olivines as Positive - Electrode Materials for Rechargeable Lithium Batteries. J.Electrochem. Soc 144, 1188 1194 (1997). 3. Nytén, A., Abouimrane, A., Armand, M., Gustafsson, T. & Thomas, J. O. E lectrochemical performance of Li2FeSiO4 as a new Li - battery cathode material. Electrochem. commun. 7, 156 160 (2005). 4. Guo, H. et al. batteries. Electrochim. Acta 55, 8036 8042 (2010). 5. Cho i, D. et al. LiMnPO4 nanoplate grown via solid - state reaction in molten hydrocarbon for Li - ion battery cathode. Nano Lett. 10, 2799 805 (2010). 6. Weker, J. N., Li, Y., Shanmugam, R., Lai, W. & Chueh, W. C. Tracking Non - Uniform Mesoscale Transport in LiFeP O 4 Agglomerates During Electrochemical Cycling. ChemElectroChem (2015). doi:10.1002/celc.201500119 7. Xu, B., Qian, D., Wang, Z. & Meng, Y. S. Recent progress in cathode materials research for advanced lithium ion batteries. Mater. Sci. Eng. R 73, 51 65 ( 2012). 8. Nanjundaswamy, K. S. et al. Synthesis , redox potential evaluation and electrochemical characteristics of NASICON - related - 3D framework compounds. Solid State Ionics 92, 1 10 (1996). 9. Padhi, A. K., Nanjundaswamy, K. S., Masquelier, C. & Goodenou gh, J. B. Mapping of Transition Metal Reclox Energies in Phosphates with NASICON Structure by Lithium Intercalation. J. Electrochem. Soc. 144, 2581 2586 (1997). 10. Masquelier, C., Padhi, A. K., Nanjundaswamy, K. S. & Goodenough, J. B. New Cathode Material s for Rechargeable Lithium Batteries: The 3 - D Framework Structures Li 3Fe2(XO4)3(X=P, As ). J. Solid State Chem. 228 234 (1998). 11. Gopalakrishnan, J. & Rangan, K. K. V2(PO4)3: A Novel NASICON - Type Vanadium Phosphate Synthesized by Oxidative Deintercalati on of Sodium from Na3V2(O4)3. Chem. Mater. 4, 745 747 (1992). 12. Gover, R., Bryan, A., Burns, P. & Barker, J. The electrochemical insertion properties of sodium vanadium fluorophosphate, Na3V2(PO4)2F3. Solid State Ionics 177, 1495 1500 (2006). 70 13. Palomar es, V., Casas - Cabanas, M., Castillo - Martínez, E., Han, M. H. & Rojo, T. Update on Na - based battery materials. A growing research path. Energy Environ. Sci. 6, 2312 (2013). 14. Lim, S. Y., Kim, H., Shakoor, R. A., Jung, Y. & Choi, J. W. Electrochemical and Thermal Properties of NASICON Structured Na3V2(PO4)3 as a Sodium Rechargeable Battery Cathode: A Combined Experimental and Theoretical Study. J. Electrochem. Soc. 159, A1393 A1397 (2012). 15. Chihara, K., Kitajou, A., Gocheva, I. D., Okada, S. & Yamaki, J. Cathode properties of Na3M2(PO4)2F3 [M = Ti, Fe, V] for sodium - ion batteries. J. Power Sources 227, 80 85 (2013). 16. Barpanda, P. et al. Sodium iron pyrophosphate: A novel 3.0V iron - based cathode for sodium - ion batteries. Electrochem. commun. 24, 116 119 (2012). 17. Palomares, V. et al. Na - ion batteries, recent advances and present challenges to become low cost energy storage systems. Energy Environ. Sci. 5, 5884 (2012). 18. Slater, M. D., Kim, D., Lee, E. & Johnson, C. S. Sodium - Ion Batteries. Ad v. Funct. Mater. 23, 947 958 (2013). 19. Huili, P., And, H. Y. - S. & Liquan, C. Room - Temperature Stationary Sdoium - Ion Batteries for Large - Scale Electric Energy Storage. Energy Environ. Sci. (2013). doi:10.1039/C3EE40847G 20. Thackeray, M. M. Spinel Electro des for Lithium Batteries. J. Am. Ceram. Soc 82, 3347 3354 (1999). 21. Shenouda, A. Y. & Murali, K. R. Electrochemical properties of doped lithium titanate compounds and their performance in lithium rechargeable batteries. J. Power Sources 176, 332 339 (20 08). 22. Nakahara, K., Nakajima, R., Matsushima, T. & Majima, H. Preparation of particulate Li4Ti5O12 having excellent characteristics as an electrode active material for power storage cells. J. Power Sources 117, 131 136 (2003). 23. Yao, X. L. et al. Comp arisons of graphite and spinel Li1.33Ti1.67O4 as anode materials for rechargeable lithium - ion batteries. Electrochim. Acta 50, 4076 4081 (2005). 24. Xiao, X., Lu, J. & Li, Y. LiMn2O4 microspheres: Synthesis, characterization and use as a cathode in lithium ion batteries. Nano Res. 3, 733 737 (2010). 25. Mumme, W. G. & Reid, A. F. Non - stoichiometric Sodium Iron Titanate, Acta Crystallogr. B 24, 625 631 (1968). 71 26. Moring, J. & Kostiner, E. The Crystal Structure of Na4Ni7(PO4)6. J. Solid State Chem. 62, 105 111 (1986). 27. Kuhn, A. et al. Study of the Conductivity of Nax - del FexTi2 - xO4 (x=0 . 875 ,del<0.40 ). J. Solid State Chem. 173, 168 173 (1998). 28. - . J. Sol - Gel Sci. Technol. 6, 7 55 (1996). 29. Liu, H., Wu, Y. P., Rahm, E., Holze, R. & Wu, H. Q. Cathode materials for lithium ion batteries prepared by sol - gel methods. J. Solid State Electrochem. 8, 450 466 (2004). 30. Fu, L. J. et al. Electrode materials for lithium secondary batteries prepared by sol gel methods. Prog. Mater. Sci. 50, 881 928 (2005). 31. Hamon, Y. et al. Aluminum negative electrode in lithium ion batteries. J. Power Sources 97 - 98, 185 187 (2001). 32. Senguttuvan, P., Rousse, G., Seznec, V., Tarascon, J. - Lowest Voltage Ever Reported Oxide Insertion Electrode for Sodium Ion Batteries. Chem. Mater. 23, 4109 4111 (2011). 33. Thevenin, J. Passivation films on lithium electrodes: An approach by means of impedance spectroscopy. J. Power Sources 14, 45 52 (1985). 34. Komaba, S. et al. Fluorinated Ethylene Carbonate as Electrolyte Additive for Rechargeable Na Batteries. ACS Appl. Mater. Interfaces 3, 4165 4168 (2011). 35. Croce, F. et al. A Novel C oncept for the Synthesis of an Improved LiFePO4 Lithium Battery Cathode. Electrochem. Solid - State Lett. 5, A47 A50 (2002). 36. Huang, H., Yin, S. - C. & Nazar, L. F. Approaching Theoretical Capacity of LiFePO4 at Room Temperature at High Rates. Electrochem. Solid - State Lett. 4, A170 A172 (2001). 37. Berthelot, R., Carlier, D. & Delmas, C. Electrochemical investigation of the P2 NaxCoO2 phase diagram. Nat. Mater. 10, 74 80 (2011). 72 3. SYNTHESIS, AND ELECTROCHEMICAL CHARACTERIZATION OF P - TYPE LAYER - STRUCTURED MATERIALS : Na 2/3 [Ni 1/3 Mn x Ti 2/3 - x ]O 2 (x=0, 1/3) 3.1 Introduction 3.1.1 Ionic conduction in layered materials Layer - structured oxide materials have a general formula of A x MO 2 , where o octahedra, tetrahedra and prismatic co - ordination of the alkali atoms, and the numeric character denotes the number of alkali layers per unit cell. 1 - 3 The alkali atoms that sandwich the transition metal layers provide a shielding effect for the close - to highe r repulsive interactions and these structures have larger A - O bond length (prismatic co - ordination). The relatively large sodium atoms adopt P - type or O - type configuration while the smaller lithium atoms prefer O - type or T - type configuration environments. - layered oxides generally have higher vacancy concentration (more non - stoichiometric) than lithium - layered oxides. Non - - layered oxide containing high vacancy concentration, support fast - conduction mechanism is operative in lithium - layered oxides, as well. 4 The fast - ion mobility is these material classes is of particular interest for Na - intercalatio n electrodes as they overcome the 73 large - size effects. Hence, P - type sodium - layered materials are considerable attractive candidates as host electrodes for Na - ion batteries. 3.1.2 Literature reports on layered host materials LiCoO 2 (LCO), with O3 - layered structu re ( - NaFeO 2 prototype), has been instrumental in the commercialization of Li - ion Batteries (LIBs) for electronics applications. However, LCO is limited by high cost and moderate capacity (~140 mAh/g). Strategies for next - generation layered materials include elemental substitution to form Li[Ni 1 - x Co x ]O 2 , Li[Ni 1/3 Mn 1/3 Co 1/3 ]O 2 and similar oxides and structural stabilization to form xLi 2 MnO 3 .(1 - x)LiMO 2 to improve the electrochemical properties. 5 - 8 The electrochemistry of NaCoO 2 has been extensively studied and reversible sodium intercalation/de - intercalation has been demonstrated. 9 However, the complex interplay between Na + /vacancy ordering at the sodium sites and charge ordering in the transition metal sites results in a complex voltage profile that includes a number of steps and plateau. Formati on of various single - phase and biphasic domains is undesirable for practical applications as these often lead to poor electrochemical performance. By introducing two or more transition metal species with different valence states, it is anticipated that the complex ordering processes would get disrupted possibly resulting in solid - solution type behavior. Some of the studied quaternary oxides containing aliovalent transition metals include Na x Ni 1/3 Mn 2/3 O 2 , Na x Fe 1/ 2 Mn 1 / 2 O 2 , Na 0.45 Ni 0.22 Co 0.11 Mn 0.66 O 2 , Na 0.85 (Li 0.17 Ni 0.21 Mn 0.64 )O 2 . 10 - 15 74 3.1.3 Materials of interest This study explores the potential of titanium - based aliovalent quate rnary oxide with the composition Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 . The beneficial role of the inert titanium matrix in enhancing the structural stability of lithium quaternary oxides for cathode applications is largely known. 16 - 18 This work unravels Ti - induced stability mechanisms for sodium intercalation cathodes. Like other P2 compositions, the excellent ionic conductivity of sodium nickel titanate (P2 - NT) has been documented previou sly. 19 The presence of two different transition metals (nickel and titanium) with significantly different - couples can be select ively activated making this composition suitable as cathode and anode. Using - - - gnificantly lower capital and operating costs. Using a single material would also eliminate any cross - contamination issues between anode and cathode in manufacturing. Electronic conduction in d - block transition metal oxides occurs by polaron hopping mech anisms involving charge carrier movement between localized potential wells. 20 - 23 The relative ease of ionization and the distribution of the transition metal species determines the charge carrier concentration and the polaron hopping length, respectively. The ionization energies of Ni 2+/3+ , Mn 4+/5+ , and Ti 4+/5+ are 17, 21, and 56 eV, respectively. 24 For P2 - N T material, charge - carrier motion is anticipated to occur mostly between the nickel states since titanium has high ionization energy. However, since the nickel concentration is relatively low, this hopping process 75 might not be readily favored as seen in Figure 3 . 1 a. Inclusion of manganese in the transition metal layer to form sodium nickel manganese titanate (P2 - NMT) material would increase the intrinsic charge carrier concentration and provide additional transport pathways by forming a percolating network as shown in Figure 3 . 1 b. This chapter delves into evaluating the electrical and electrochemical properties of these novel electrode compositions. (a) (b) Figure 3 . 1 Schematic of the transition metal slab in layered oxide depicting charge carrier p athways in (a) P2 - NT , and (b) P2 - NMT through formation of percolating network; red, blue and orange s pheres denote nickel, titanium, and manganese atoms, respectively. 3.2 S tructural features 3.2.1 Sodium nickel titanate: P2 - Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 P2 - NT phase was reported to crystallize in a disordered layered structure with space group P6 3 /mmc (space group 194). 3 , 19 The studied material has the same prototype structure as the well - reported Na 0.7 CoO 2 material. 25 sequence f e site). It was reported that Na f and Na e are not energetically equivalent and the former has lower site occupancy by virtue of its larger repulsive int eractions with the transition metals. Figure 3 . 2 shows the location of two sodium layers per unit cell and hence this structure can be represented 76 by Delmas notation site. 19 Figure 3 . 2 Crystal structure of Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 ; blue spheres refer to transition metals (nickel and titanium), yellow spheres refer to sodium atoms Oxygen atoms are not included for clarity. 3.2.2 Sodium nickel manganese titanate: P2 - Na 2/3 [Ni 1/3 Mn 1/3 Ti 1/3 ]O 2 solution composition with the parent phase, P2 - NT. Hence, P2 - NMT has the same crystal structure as P2 - NT phase , shown previously in Figure 3 . 2 with random distribution of manganese. This allows for retaining the excellent sodium conduction pathways of the P2 phase while potentially improving the e lectrical transport properties in the transition metal layer. 3.3 Inorganic synthesis and high - temperature processing 3.3.1 Sodium nickel titanate synthesis The precursors were Na 2 CO 3 ( 99.5%), NiO (micron powder, 99%; <50 nm powder, 99.8%), and 77 TiO 2 (micron powder , 99%; 21 nm powder, 99.5%). For a typical synthesis of 10 g of powder, 4.6 g of Na 2 CO 3 , 2.5 g of NiO, and 5.4 g of TiO 2 were used. Na 2 CO 3 of 10 wt% excess were added to compensate the sodium oxide evaporation during high - temperature processing. Both micr on - and nano - sized NiO and TiO 2 were employed to compare the effect of particle size on the ease of synthesis and the electrochemical properties. Micron - sized precursors were subjected to dry - milling using SPEX SamplePrep 8000M mixer/mill. Nano - sized precursors were mixed in a jar mill with 2 - propanol as solvent. After uniform mixing and solvent removal at 120 o C, the powders were fired at 900 o C in a box furnace exposed to air . Figure 3 . 3 XRD characterization of sodium nickel titanate pr epared by solid state reaction from micron - sized precursors fired for 12 hours duration; The dominant peaks of the impurity and target phases are indexed. The XRD patterns of the micron - sized precursor powders fired for 12 hours duration at different tem peratures are shown in Figure 3 . 3 . Powders fired at 600 o C had considerable amount of 78 unreacted TiO 0.9 phase (ICSD reference code: 78 - 0720 ) and Na 4 Ti 5 O 12 (ICSD referen ce code: 52 - 1814 ) impurity phase along with the target Na 2/3 [Ni 1/3 Ti 23 ]O 2 phase (ICSD reference code: 4 8 - 0 164 ) . As the calcination temperature is increased to 700 and 800 o C, a reduction in the amount of unreacted TiO 2 and impurity phase is observed. It was found that heating to 900 o C is essential to obtain phase - pure powders with these precursor powders and processing conditions. When using nano - precursors, the fired powders were matched with the target phase without unreacted components as shown in Figure 3 . 4 a. even when calcined at a lower temperature of 800 o C. However, a small impurity peak was detected at 41 o and this could no t be reasonably matched with any binary or ternary Na - Ni - Ti - O oxide components in ICSD database. Attempts to eliminate this impurity by prolonging the synthesis duration to 20 and 36 hours were not successful. Completely pure target phase could only be for med at a slightly higher temperature of 900 o C when heated for at least 2 hours and further electrochemical testing was performed using these powders, as depicted in Figure 3 . 4 b. The SEM images of the as - prepared P2 - NT powders are shown in Figure 3 . 4 c and d. Synthesis from micron - sized precursors results in fairly spherical partic les that are roughly 5 - 10 µm in diameter. Nano - sized precursors yielded finer powders with particles less than 5 µm due to reduced heating duration. Nano - precursors provide pure phase powders, with smaller particle size, in shorter synthesis due to shorter diffusional length and facile reaction kinetics. 79 (a) (b) (c) (d) Figure 3 . 4 XRD characterization of P2 - NT material prepared by solid state reaction from wet - milled nano - sized precursors fired at (a) 800 o C, and (b) 900 o C, SEM morphology of powders prepared from (c) micron - sized precursors, and (d) nano - sized pre cursors. 3.3.2 Sodium nickel manganese titanate synthesis The precursors consisting of stoichiometric amounts of Na 2 CO 3 ( 99.5%), NiO (micron powder, 99%), TiO 2 (micron powder, 99%) and MnO 2 ( 99%) were dry - milled using high - energy SPEX mill and subsequently fired at 900 o C for 12 hours duration. XRD of P2 - NT and P2 - NMT phases 80 have identical patterns as both structures have the same space group (194) suggesting solid solution formation ( Figure 3 . 5 a ). The prepared P2 - NMT powders were found to be agglomerates of micron - sized particles as shown in Figure 3 . 5 b . For the purposes of this study, the valency of nickel, manganese and titanium are taken as 2+, 4+, and 4+, respectively, based on the synthetic conditions and literature reports for isostru ctural materials. 3 , 10 , 19 (a) (b) Figure 3 . 5 (a) XRD comparisons of sodium nickel titanate and sodium nickel manganese titanate powders prepared by solid state reaction, (b) and SEM powder morphology. T he surface area of the as - synthesized powders were measured using the B runauer - E mmett - T eller (BET) method with Micromeritics ASAP 2020 instrument employing krypton as the adsorbate gas. The powders were subjected to a degassing step where the powders were preheated in a sample holder under low - pressure co nditions to eliminate any physisorbed water and volatile adsorbed components. The BET surface area of the P2 - NT (micron - sized precursors), P2 - NT 81 (micron - sized precursor) and P2 - NMT (micron - sized precursor) powders were estimated to be 0.58, 1.30 and 0.78 m 2 /g, respectively. This translates to an average particle size of 1.33 µm and 0.97 µm for P2 - NT and P2 - N M T, respectively , which is used subsequently for diffusion co - efficient calculations . 3.3.3 Powder sintering The as - p repared oxide powders were cold - pressed a t a pressure of 2 MPa and sintered at a temperature of 1000 o C for 10h to obtain a dense pellet. The surface morphology of the P2 - NT and P2 - NMT pellets are depicted in Figure 3 . 6 . Both pellets appear to have sintered without any visible surface pores. Quantitative estimates of t he pellet densities were obtained using the Archimedes method with 2 - Propanol as solvent. The P2 - NT and P2 - NMT pellets were measured to have relatively high density values of 93 and 89%, respectively, substantiating the SEM - morphology observations. The sintered pellets were pasted with gold contacts on both sides for further electrical measurements (DC and impedance spectroscopy testing). The s intered pellets were 2.3 mm thick and 12.6 mm in diameter. (a) (b) Figure 3 . 6 (a) Surface morphology of sintered pellets of (a) P2 - Na 2/3 [Ni 1/3 Ti 2/3 ]O 2 , and (b) P2 - Na 2/3 [Ni 1/3 Mn 1/3 Ti 2/3 ]O 2 . 82 3.4 F ilm fabrica tion The previously established protocols (chapter 2) were followed for fabricating the porous, composite electrode films with P2 - NT and P2 - NMT as active components. The film contains 80:10:10 weight ratio of active component, polymeric binder and conductive ca rbon. The punched film electrodes were transferred to a glovebox to reduce exposure to atmospheric water vapor and carbon dioxide. 3.5 Material Characterization 3.5.1 Voltage profile under galvanostatic testing 3.5.1.1High - voltage cathode: Ni 2+/4+ redox couple Sodium de - intercalation/intercalation in P2 - NT would initiate the Ni 2+/4+ redox couple and the theoretical capacity is estimated to be 181 mAh/g for complete sodium content range The redox reactions can be represented as 3 . 1 P2 - NMT would have identical reaction as Equation 3 . 1 in the cathodic regime activating only the nickel redox couple with manganese remaining inert at valency of 4+ in the studied voltage range. The theoretical capacity of this phase is calculated to be 176.6 mAh/g assuming complete sodium exchange. 83 (a) (b) (c) (d) Figure 3 . 7 Cathodic voltage profile under galvanostatic testing rate of C/50 rate for (a) P2 - NT material with cutoff of 4.2 V and, (b) P2 - NT material with cutoff of 4.5 V, (c) P2 - NMT ma terial with cutoff of 4.2 V and, (d) P2 - NMT material with cutoff of 4.5 V. Figure 3 . 7 illustrates the voltage profile of P2 - NT and P2 - NMT phases under galvanostatic conditions of C/50 rate. In P2 - NT, the first charge/discharge curve has a large irreversible capacity 84 and substa ntial voltage hysteresis due to electrolyte side reactions. 26 During the subsequent cycles, the voltage gap gets significantly diminished and the material exhibits good reversibility. The voltage profile appears sloping and does not involve multiple steps and plateau indicat ive of solid solution reaction mechanism when cycled till 4.2 V. However, a significant voltage hysteresis is noticeable upon increasing the charging voltage to 4.5 V. Additionally, a voltage plateau is observed at around 4.2 V corresponding to possibly a two phase reaction. Operating in the multiphasic domains contributes to poor reversibility as seen from Figure 3 . 7 b. P2 - NMT exhibits excellent reversible sodium intercalation/de - intercalation reactions as illustrated in Figure 3 . 7 c. 27 The voltage curves are sloping with very small voltage gap and good faradaic efficiency. Unlike P2 - NT, the first charge/discharge cycle does not have a significant irreversible capacity showing that P2 - NMT does not favor electrolyte oxidation as much as P2 - NT . With high charging cutoff voltage of 4.5 V, a voltage plateau appears at around 4.2 V and this contributes to large voltage hysteresis and poor reversibility Figure 3 . 7 d). The high redox potential (E o ~3.7 V vs. Na/Na + ) of these materials makes them attractive as high - voltage sodium intercalation cathodes. 3.5.1.2 Low - voltage anode: Ti 4+/3+ redox couple Titanium redox couple in P2 - NT is activated by filling - in the v acant sodium sites in the pristine composition. This is made feasible by the presence of unusually high number of of vacant sites in this non - stoichiometric oxide. The sodium intercalation reaction can be represented by Equation 3 . 2 . 3 . 2 85 Figure 3 . 8 Anodic voltage profile of sodium nickel titanate under galvanostatic testing of C/5 rate with 0.2 V cutoff voltage. During the 1 st cycle charging curve, the amount of intercalated sodium is higher than the theoretical capaci reductive potentials (E~0.2 V vs. Na/Na + ), as shown in Figure 3 . 8 . It is anticipated that these degradation reactions are promoted by conductive carbon additive and P2 - NT active particles. Nevertheless, the intercalated sodium ions could be extracted in the subsequent discharge cycle indicating that these degradation reactions do not form a blocking surface film. During the subsequent cycles, very little side reaction is observed with full realization of the theoretical capacity in the studied voltage window. The voltage gap between the charge and discharge curves is relatively small even at high charging currents of C/5 indicating that this composition has excellent reversibility in the anodic regime. Ti 4+/3+ redox couple was found to have a low reduction potential (E o =0.7 V vs. Na/Na + ) and hence this phase shows promise as low - voltage, intercalation - based anode. 86 P2 - NT has been shown to demonstrate reversible sodium intercalation/de - intercalation reactions using Ni 2+/4+ and Ti 4+/3+ redox couples with E o values of 3.7, and 0.7 V, respectively. This composition can be utilized as - - for the design of 3 V battery devices using the same material as anode and cathode and is anticipated to provide substantial cost - reduction in capital and operating costs by switching over - - contamination issues. 3.5.2 Quantifying charge carrier transport processes: DC & AC Testing The electrical response of mixed conducting pellets with ion - blocking metal electrodes (gold) can be represented using a circuit model that accounts for the two charge carrier transport mechanisms, i.e. electronic and ionic, as illustrated in Figure 3 . 9 a. The ionic conduction rail has circuit elements to account for the resistance to ion hopping within individual grains (R ion ), resistance to ion hopping across grains with different orientati ons (R gb ), capacitive contribution at the grain boundary (C gb ) and intercalation capacitance (C int ). The electronic conduction rail is simplified to a single resistive element with ohmic response (R eon ). A capacitive element accounting for the pellet geome tric capacitance effects (C geom ) by passes the electronic and ionic conduction rails. 28 Given that the tested pellets are reasonably thick (2.3 mm), the geometric capacitance rail is ignored as it correlates inversely with the separation distance as given by the expression: r o - At the high - frequency limit, the interfacial capacitance and grain boundary capacitance become shorted as the impedance of capacitive elements correlate inversely with AC signal frequency as 87 given by the expression: c refers to the angular frequency. Hence, the equivalent circuit gets reduced to Figure 3 . 9 b. At the reduced to just the electronic conduction rail as shown in Figure 3 . 9 c. (a) (b) (c) Figure 3 . 9 (a) Electrical equivalent circuit model representing physical processes in a mixed electronic - ionic conductors with blocking electrodes, (b) Limiting model for high - frequency AC signal, and (c) Limiting model for DC signal. 28 The total conductivity due to all charge carriers (ions and electrons) is estimated from the high freque ncy intercept of the impedance curves. For the impedance measurement, the impedance curves gets smaller (total conductivity increases) with rising temperature due to an overall enhancement of charge carrier transport processes as shown in Figure 3 . 10 a. For DC testing, Figure 3 . 10 b. shows the measured current response after a pplying a step potential of 200 mV . There is an initial capacitive decay and steady state current is attained at longer durations that is related to the steady - state supply of charge carriers . The data is fitted with an exponential curve to extrapolate the steady state current and subsequently obtain electronic conductivity values at different temperatures . 88 (a) (b) Figure 3 . 10 Electrical testing of P2 - NMT sintered pellet: (a) High - frequency AC impedance spectra in temperature range of 383 - 423 K with 10 K increments, and (b) Current decay response upon applying a potential of 200 mV at a 383 K. The total conductivity curves and the electronic conductivity curves obey Arrhenius temperature relation ( Figure 3 . 11 ). T he total conductivity is around 4 orders magnitude higher than the electronic conductivit y for both phases. This shows that sodium ions are the dominant cha rge carrier s for these layered materials . The electronic conductivity of P2 - NMT (1.21 × 10 - 7 S/cm at 110 o C) is higher than P2 - NT (3.96 × 10 - 8 S/cm at 110 o C) with E a values of 0.574 eV and 0.862 eV, respectively. The charge carrier transport is demonstrated to show a significant improvement by manganese incorporation by providing a percolation network in the transition metal layer. It is also found that P2 - NMT has total conductivity of 8.28 × 10 - 3 S/cm (at 110 o C) with activation energy (E a ) of 0.279 eV while P 2 - 89 NT has total conductivity of 4.89 ×10 - 3 S/cm (at 110 o C) with E a of 0.296 eV. M anganese addition improves ion transport rate by a factor of 1.7 . Figure 3 . 11 Total effective conductivity (ionic and electronic) and electronic conductivity curves as a function of temperature from AC impedance and DC testing. Nevertheless, the intrinsic electronic conductivity values of both materials are still relatively low and this needs to be addressed (using stra tegies such as aliovalent doping, elemental substitution, reducing the particle size etc.) to achieve improved electron transport and subsequently better electrochemical properties at high currents. As the conductivities curves follow an Arrhenius correlat ion and the conductivity values are small, it is likely that there is only a small overlapping of the conduction bands and charge transport is facilitated by hopping of small polarons between localized potential wells ( between the 3d transition metal speci es ). A similar conductivity trend was reported for P3 - Na 0.6 [Ni 0.6 Co 0.4 ]O 2 but with much higher electronic 90 e >10 - 2 S/cm at 270 K) facilitated by the faster electron hopping via cobalt and nickel atoms . 29 3.5.3 E valuating ionic transport and interfacial processes: Electrochemical impedance spectroscopy Electrochemical impedance spectroscopy (EIS) is utilized as a quantitative tool to isolate various sodium - ion transport processes in the composite electrode that have widely separated time constants. It has been demonstrated that the basic electrical circu it elements such as resistors, capacitors and their modified combinations have similar mathematical formulation as physical, chemical and electrochemical processes. For instance, the ion transport in electrolytes can be represented by Equation 3 . 3 which takes the same mathematical form as an ohmic resistor 3 . 3 i i i i i i is the ohmic resistance. Interfacial charge transfer reactions can be represented by a resistive circuit element (referred o y Equation 3 . 4 . 3 . 4 91 These circuit elements are applied for emulating the multi - time scale physicochemical processes and interpreting i mpedance spectra data. Assigning different parameters to various physicochemical processes permits identification of the bottleneck processes. However, several circuit models can provide a good match to the experimental spectra and the selected model need s to have reasonable physical meaning for accurate interpretation. This requires a priori knowledge of the basic processes for a given system and state - of - the - art models for similar materials. Figure 3 . 12 Electrical equivalent circuit model depicting sodium ion transport processes (within liquid electrolyte, across surface films, across the interface, and bulk solid state diffusion) of the composite electrode; where Re is the electrolyte resistance, Rsf is the surface film resistance, CPE sf is the constant phase element due to surface films, CPE ct is the constant phase element due to the double layer, and Z w is the semi - infinite Warburg diffusional element. The impedance curves of Na - intercalation electrod es are analyzed using an analogous circuit model developed previously to i nterpret lithium intercalation phenomenon in graphite and lithium transition metal oxides 30 - 32 as depicted in Figure 3 . 12 . Nyquist plots of P2 - NT and P2 - NMT electrodes display the following features: two ove r - lapping semi - circles in the high frequency domain, one semi - circle in the medium frequency domain and a small diffus ion tail at the low frequency domain (depending on the electrode potential) as seen in Figure 3 . 13 . The semi - circles at the high and medium frequency region are attributed to the relatively fast transport of sodium 92 ions through the surface films at the film/electrolyte interface (at active material and sodium metal), coupled with film capacitance. These surface films can be formed on the pristine material due to interaction with atmospheric CO 2 or decomposition of electrolyte constituents during sodium intercalation/de - intercalation and have been reporte d for intercalation materials. 30 The capacitance has been modeled using a constant phase element (CPE) to account for the depressed semi - circles formed possibly due to surface heterogeneities. The effective capacitance (C) is calculated using the re lation, C= Q R - 1) ; 33 , 34 are the CPE fitting parameters The low frequency semi - circle is assigned to the interfacial charge transfer of sodium ions (R ct ), coupled with capacitance (C ct ) at the electrode/surface film interface. The analysis of the impedance spectra was performed using Zview impedance analysis so ftware (Scribner Associates Inc., Southern Pines, NC). Non - linear Least Square fitting (NLLS) provides a good fitting of the experimental data as seen in Figure 3 . 13 . The linear tail region in the low frequency region is attributed to diffusion phenomenon discussed subsequent ly . The C ct is measured to be in the order of few mF, much higher than the film capacitance (10 - 4 - 10 - 6 F) as seen in Figure 3 . 14 a . The true surface areas of the porous electrode particles are much higher than the geometric area and hence we get these high cap acitance values. The film capacitance values are in t he range of µF as they correspond to surface film/electrolyte interface as seen in Figure 3 . 14 b. 93 (a) (b) (c) Figure 3 . 13 Nyquist plot of P2 - NT (blue) and P2 - NMT (yellow) electrodes scanned between frequency range of 5 MHz to 10 mHz with inset showing high - frequency depressed semicircles at voltage of (a) 3.4 V vs. Na/Na + , (b) 3.8 V vs. Na/Na + , and (c) 4.2 V vs. Na/Na + . 94 (a) (b) Figure 3 . 14 Potential dependence of (a) charge transfer resistance and double layer capacitance, and (b) surface film resistance and surface film capacitances. Figure 3 . 14 a shows that the interfacial charge transfer resistance (R ct ), for both P2 - NT and P2 - NMT, has a strong dependence on electrode potential and decreases by several orders of magnitude as the electrode potential increases from 3.0 to 4.2 V. At highly oxidative potentials, the increased amount of Ni 4+ /Ni 2+ states allows for enhanced mobility of the charge carriers (between these multivalent states) resulting in faster interfacial kinetics. Similar strong correlation of interfacial kinetics with electrode potential has been previously observed for Li 1 - x CoO 2 electrodes. 31 Also, it is found that the R ct of P2 - NMT is c onsiderably smaller than P2 - NT over the complete range of measured electrode potentials indicating that the former mater ial supports faster interfacial transfer than the latter . This trend aligns with the polaron attributed electronic conductivity boasting mechanism discussed previously. The associated double - layer capacitance values (C ct ) remains fairly independent of the e lectrode potentials and is comparable for both materials. This is 95 anticipated as the electrode surface area does not change significantly during sodium de - intercalation. The electrolyte resistance is invariant with electrode potential as the concentration of sodium ions is constant. The surface film resistances (R sl1 and R sf2 ) and surface film capacitances (C sf1 and C sf2 ) are largely invariant with electrode potential suggesting that these films are fairly stable and do not change significantly in their chemical make - up/composition (that can affect the dielectric constant) and film thickness in the high voltage region. The solid state diffusion of sodium ions within the active particles contributes to the Warburg response in the low frequency domain as se en by the characteristic linear response in the Nyquist plots (45 o tails) depicted in Figure 3 . 13 b and c (seen distinctly for P2 - NMT). We make use of the conventional planar diffusion model, typically used to simulate intercalation electrodes, to interpret the experimental data. 31 , 35 The imaginary and real components of the impedance can be expressed as Equation 3 . 5 and Equation 3 . 6 : 3 . 5 3 . 6 j real m is the molar volume, is the slope of the voltage - is the Faraday consta is the chemical diffusion co - is the angular frequency, w - factor. 96 The Warburg pre - factor (A W ) is estimated from linear fitting of the experimental data as shown in Figure 3 . 15 The average of A w from the real and imaginary part of impedance was used to obtain the apparent chemical diffusion coefficient of sodium ions in the host materials. The impedance curves of P 2 - NT (all measured voltages) and P2 - NMT at 3.0 V do not show a distinctive diffusion tail (within the measured frequency domain) and hence are not considered for further analysis. This is because the slow interfacial kinetics overshadows the diffusion proc ess. P2 - NMT, from EIS, are reasonably high (Mean D=9.1× 10 - 13 cm 2 /s), as typical of intercalation electrodes . Further discussion of the diffusion co - efficients is provided in the subsequent electroanalytical section. Figure 3 . 15 - 0.5 curves using planar diffusion model to obtain Warburg prefactor. 97 3.5.4 Analyzing ionic transport processes: Potentiostatic intermittent titration technique P otentiostatic Int ermittent Titration technique (P ITT ) provides a complimentary method to evaluate the diffusional properties of the studied materials. Low limitation currents of C/100 were imposed to approach close - to - equilibrium conditions. The electrodes were subjec ted t o initial precycling at C/10 rate before the PITT experiments ( as depicted in inset of Figure 3 . 16 a and b ). The voltage profiles of P2 - NT and P2 - NMT materials are fai rly comparable to galvanostatic testing . P2 - NT has a solid solution behavior evidenced by the sloping voltage trend when cycled with voltage limit of 4.2 V. It has a low faradaic efficiency as the low limiting current values feed the irreversible side reac tions. P2 - NMT has sloping voltage profile till 4. 1 V after which a plateau is noticed. The PITT voltage profile of P2 - NT material in the anodic regime ( Figure 3 . 16 c) has significantly large irreversible capacity loss compared to the galvanostatic testing. The increase in side reactions at slower charging rate suggests that these side reactions are catalyzed chemically at these reductive potentials rather than electroc hemical mechanisms. Long - time approximation method is used to extract diffusional information from PITT measurements . Kinetic limitations, that are noticeable in the impedance analysis, are overcome by using a relatively long relaxation time (current drops to C/100) between the incremental steps. The expression for current transient upon impressing a constant potential is given by Equation 3 . 7 . 3 . 7 potential step, is the chemical diffusion co - efficient, L is the diffusion length scale, and t is the time taken for relaxation. 98 (a) (b) (c) Figure 3 . 16 Voltage profile under potentiostatic intermittent titration technique using C/100 current limitation for (a) cathodic P2 - NT with 5 mV voltage step, and (b) cathodic P2 - NMT with 5 mV voltage step, and (c) anodic P2 - NT with 15 mV voltage step; inset in cathodic curves depict electrode precycling at C/10 rate . 99 (a) (b) (c) Figure 3 . 17 (a) Linear fitting of current transients from potentiostatic intermittent titration technique at different electrode potentials with current limitation of C/100 for cathodic P2 - NT electrode, (b) linear fitting of P2 - NMT elect rode under identical conditi ons, and (c) variation of apparent chemical diffusion co - efficient with electrode potential as calculated from potentiostatic intermittent titration technique and impedance spectroscopy. 100 Figure 3 . 17 a and b displays the linearity of the experimental data for both NT and NMT materials validating the use of this PITT model and the slope is used to estimate the diffusion co - efficients. The e - 12 - 10 - 13 cm 2 /s) and comparable to EIS estimates. The diffusivity of P2 - NMT increases with voltage and falls off at 4.2 V as evident from the PITT and EIS estimations in Figure 3 . 17 c. P2 - NT also follows a similar rising trend with voltage. This can be rationalized by the fact that sodium diffusivity increases due to the creation of additional sodium vacancies. However, i t is likely that the low concentration of sodium ions charge carriers leads to diffusivity drop at high voltages. The diffusivity dip of P2 - NMT also gets convoluted by the presence of additional phases that form at high charging voltage leading to breakdow n of current - transient model assumptions. 3.5.5 Understanding reaction mechanisms: Ex - situ XRD Testing Ex - situ XRD testing on the P2 - NT film electrodes at different stages of charging shows that the material retains its P2 - layered structure during sodium removal as shown in Figure 3 . 18 a . The o starts to shift towards lower angles indicating that the structure starts to expand along the c - direction (i nset in Figure 3 . 19 a) . As sodium content decreases in these P2 materials, the oxygen atoms in the adjacent layers start experiencing increased repulsive interactions c ontributing to the expansion along the c - axis. A solid - s olution reaction mechanism till charg ing voltage of 4.2 V without any phase change is confirmed by the XRD results. However, when the charging voltage is increased to 4.5 V, an unidentified phase poss ibly related to O - type structure forms ( Figure 3 . 18 b) and this phase change responsible for poor reversibility in the high voltage region . 101 (a) (b) (c) (d) Figure 3 . 18 Ex - situ XRD of electrode films: (a) sodium nickel titanate cathode charged till 4.2 V with inset showing the evolution of 16o reflection, (b) sodium nickel titanate cathode charged to 4.5 V, (c) sodium nickel manganese titanate cathode charged till 4.2 V, and (d) sodium nickel titanate anode charged to 0.2 V with inset showing evolution of 16 o reflection; all potentials are measured with respect to Na/Na + . 102 P2 - NMT also demonstrates a solid - solution mechanism till charging voltage of 4.0 V where the pristine phase peaks are retained as seen from Figure 3 . 18 c . Removing further sodium leads to format ion of a low - angle peak at 12.5 o . During s ubsequent sodium intercalation, this peak vanishes and the film returns to its initial state sug gesting good reversibility. This is possibly due to sodium ordering processes along the c - direction . P2 - NT electrode in the anodic regime appears fairly stable without the formation of any metallic components (such as nickel, titanium) confirming that the low - voltage reactions are due to intercalation mechanism as shown in Figure 3 . 18 d. Very little change in peak positions (16 o ) suggests that the lattice parameters of the fully intercalated phase is fairly comparable to that of the pristine phase. The XRD pattern of th e cycled electrode also remains identical to the pristine phase suggesting that the material does not degrade significantly during the electrochemical cycling. 3.5.6 Evaluating capacity retention 3.5.6.1 Cycling Study Capacity retention during repeated charge/ discharge cycles is essential to designing long - life battery devices. However, it is found that the repeated intercalation/de - intercalation processes contributes causes mechanical strain effects as a result of host - intercalant interactions leading to capac ity fade. Other fade mechanisms include electrolyte degradation, dissolution of active materials, defect formations, etc. 36 The cycling study in this research work is limited to half - cell testing that poses inherent limitations due to sodium dendritic formation. In the cathodic regime, the faradaic efficiency of the P2 - NT phase is poor in the initial few cycles and it improves subsequent cycling. The discharge capacity is found to be fairly stable 103 during the tested 30 cycles as shown in Figure 3 . 19 a P2 - NMT phase has an initial stabilization period of 10 cycles after which it shows excellent charge retention with no loss in discharge capacity between 11 th and 50 th cycle. Figure 3 . 19 b illustrates that the faradaic efficiency of NMT phase is fairly close to 100%. The initial stabilization phase is attributed to electrolyte side reactions. The exce l lent capacity retention of the P2 materials in cathodic regime , even in half - cell testing, is due to the very small volume expansion as reported for similar structures. 37 Since the material does not undergo significant str ain during sodium exchange processes , the active particles remain fairly intact and contribute to good cycling performance . These compositions show promise for designing long - life, intercalation host electrodes. In the anodic regime, P2 - NT phase shows a continuous capacity fade as observed in Figure 3 . 19 c. Considering that the XRD patterns of the cycled elect rode remained identical to the pristine phase, this capacity fade is not associated with bulk structural changes that would lead to loss of crystallinity. We predict that this is due to electrolyte degradation at the low potentials that gets catalyzed by n ickel atoms in this P2 composition. Recently a lithiated P2 phase with identical prototypical structure has been shown to have excellent capacity retention. 37 Future work would need to focus on identifying electrolyte degradation components using spectroscopic characterization techniques. This would isolate the solvent component that is responsible for the fade mechanism and enable development of better electrolyte co mbinations. 104 (a) (b) (c) Figure 3 . 19 Capacity retention study with 4.2 V upper voltage cutoff for (a) cathodic sodium nickel titanate at C/10 rate, (b) cathodic sodium nickel manganese titanate electrodes tested at C/5 rate, and (c) anodic sodium nickel titanate tested at C/5 rate. 3.5.6.2 Impedance change during cycling Impedance spectra of P2 - NMT phase was analyzed during the cycling test after the charging the electrodes to 4.2 V as shown in Figure 3 . 20 a. The curves remain fairly consistent suggesting that the underlying physicochemical processes are identical for the pristine and cycled electrode. 105 (a) (b) (c) Figure 3 . 20 (a) Nyquist impedance spectroscopy plot of sodium nickel manganese titanate electrode scanned in frequency range of 5 MHz to 10 mHz after charging to 4.2 V, (b) variation of imp edance resistive parameters, and (c) capacitive parameters during cycling . 106 The electrolyte resistance, from the high - frequency intercept, was found to be invariant during the electrode cycling as shown in Figure 3 . 20 b. Even though the electrolyte composition is not optimized and sodium metal still reacts with the solvent, those effects did not affect the ionic conductivity considerably. The interfacial ch arge transfer resistance, after the initial stabilization period, remains invariant. The surface film resistances show only minor fluctuations during the testing. The invariant double - layer capacitance can be correlated to the stable surface area of the ac tive particles, while the invariant surface film capacitance correlates to insignificant change to the dielectric properties of the surface films ( Figure 3 . 20 c). Thes e suggests that the side reactions get arrested after the initial cycles and a relatively stable passivation film (SEI) gets formed at the electrode/electrolyte interface. 3.5.7 Rate performance testing Cathodic P2 - NT material prepared from nano - sized precursors provides capacity values of 81 mAh/g, 77 mAh/g, 68 mAh/g, 34 mAh/g, and 9 mAh/g at C/20, C/10, C/5, 1C, and 2C rates, respectively, as shown in Figure 3 . 21 a. Cathodic P2 - NT material prepared from micron - sized precursors yielded capacity values of 60 mAh/g, 53 mAh/g, 45 mAh/g, 17 mAh/g and 3 mAh/g at C/20, C/10, C/5, 1C and 2C, respectively. The cathodic P2 - NT powders with smaller particle size delivers highe r capacity values due to shorter charge - carrier transport length scales. It was also found that P2 - NMT material (micron - sized precursors) provides capacity values of 90 mAh/g, 80 mAh/g, 65 mAh/g, 54 mAh/g and 40 mAh/g at C/20, C/10, C/5, 1C and 2C, respe ctively ( Figure 3 . 21 b). Under identical processing and calcination conditions, P2 - NMT phase clearly has improved capacity retention and is more suitable for high powe r applications. As discussed in the previous sections, this capacity boast with manganese substitution is attributed to accelerated interfacial kinetics of sodium ion transfer at the electrode/electrolyte interface, 107 relatively unhindered electron hopping f acilitated between manganese and nickel sites, and higher diffusivity of sodium ions as demonstrated from EIS, DC measurements and PITT analysis. (a) (b) (c) Figure 3 . 21 High - current rate testing of (a) cathodic sodium nickel titanate powders with micron - and nano - sized powders, (b) cathodic sodium nickel titanate and sodium nickel manganese titanate electrodes with 4.2 V cutoff, and (c) anodic sodium nickel titanate wi th cutoff voltage of 0.2 V. . Anodic P2 - NT furnishes capacity values of 75 mAh/g, 73 mAh/g, 70 mAh/g, 53 mAh/g, 43 mAh/g, and 30 mAh/g at charging rates of C/20, C/10, C/5, 1C, 2C, and 4C respectively ( Figure 108 3 . 21 c). These values are significantly better than those in the cathodic region for the same composition. The improved rate performance in the anodic region is due to the presence of large number of titanium atoms, i.e. 2/3 of Ti compared to 1/3 of Ni, forming a percolating network leading to enhanced electron conduction (due to Ti 4+/3+ ) in the transition metal layer. 3.6 Improving cathodic stability: Calcium substitution During charging, the sodium content becomes reduced in the layered host materials and this leads to excessive unfavorable repulsive interactions and phase change. This limits the electrochemical window and also leads to loss of reversible capacity. The scope of this preliminary study was to potentially stabilize the layered ox ides using calcium substitution. Calcium is anticipated to occupy the sodium layer and provide structural stability at high charging voltages. (a) (b) Figure 3 . 22 (a) XRD characterization of 5% and 20% calcium substituted sodium nickel titanate prepared by solid - state reaction at 900 o C, and (b) EDS mapping of 5% sodium nickel titanate powders to check calcium distribution. 109 Sodium was partially replaced with calcium to prepare target compositions of [Na 0.53 Ca 0.07 ][Ni 0.33 Ti 0.67 ]O 2 (20% calcium substituted) and [Na 0.63 Ca 0.02 ][Ni 0.33 Ti 0.67 ]O 2 (5% calcium substituted). The stoichiometric precursors consisting of Na 2 CO 3 , nano - NiO, nano - TiO 2 and CaO (99.9%) were wet milled as discussed before and fired at 9 00 o C. The XRD patterns of 20% Ca - substituted material indicated that the calcium did not get incorporated in the layered structure and formed a perovskite secondary phase with composition CaTiO 3 ( ICSD reference code: 88 - 0790) as seen in Figure 3 . 22 a. The powders prepared with a lower 5% calcium content also displayed the characteristic perovskite peaks. EDS mapping was performed on the 5% calcium - powders to check the overall distribution of calcium as seen purple spots in Figure 3 . 22 b. Nickel distribution is overlayed as an internal standard. The calcium distribution is found to be fairly uniform without significant enrichment due to perovskite particles. This s uggests the possibility of low - level calcium substitution or perovskite phase forming a thin coating on all the particles. Since we did not have a solid evidence for calcium forming a solid solution, these powders were not electrochemically characterized. Nevertheless, a recent research study supports the concept of calcium - attributed stabilization in isostructural Na 0.7 CoO 2 with significantly improved capacity retention. 38 3.7 Improving anodic capacity: Potassium substitution The anodic capacity of the P2 - NT phase is limited by the number of available vacant sites in the structure; 1 mole of Na 2 /3 [Ni 1/3 Ti 2/3 ]O 2 theoretical capacity to 90 mAh/g. Recently, a lithiated P2 phase, Na 2/3 [Li 1/3 Ti 2/3 ]O 2 , was found to have excellent electrochemical properties in the anodic regime. 37 For these P2 non - stoichiometric 110 materials, even a modest increase in vacancy concentration would provide a significant boast to the sodium intercalation capacities as seen from Table 3 . 1 . Table 3 . 1 Theoretical capacity of P2 - titanate materials based on vacancy concentration. Composition Vacancy (1 - x) Theoretical capacity (mAh/g) Molar mass (g/mol) Na 0.67 [Li 0.33 Ti 0.67 ]O 2 0.33 105.9 86.0 Na 0.55 [Li 0.18 Ti 0.82 ]O 2 0.45 141.9 85.0 Na 0.5 [Li 0.17 Ti 0.83 ]O 2 0.50 158.5 84.5 (a) (b) Figure 3 . 23 XRD characterization of powders to synthesize (a) sodium deficient compositions with x=0.67, 0.55 and 0.50 at firing temperature of 1000 o C for 20 h duration, and (b) Na 2/3 - y K y [Li 1/3 Ti 2/3 ]O 2 for y=0, 0.05, 0.10 and 0.15 prepared at firing temperature of 800 o C for 20 h. Solid - state synthetic technique was utilized to prepare the aforementioned low sodium content phases with theoretical capacity of ~158 mAh/g, i.e. 75% higher than the pristine phase. The pristine lithiated phase (x=0.67) was successfully prepared with soli d - state synthesis technique by firing the precursor powders at 1000 o C for net duration for 20 hours; the calcined powders were 111 ground after heating for 10 hours to promote powder homogeneity and fired for additional 10 hours. Figure 3 . 23 a confirms that the prepared powders of the pristine phase were phase - pure. Sodium deficient phases (x=0.55 and x=0.50) were prepared by identical methods but resulted in the formation of [NaLi]Ti 3 O 7 impurity phase (ICSD reference: 52 - 0690). This is due to structural destabilization due to increased alkali layer spacing. Samples were prepared with potassium contents of 0, 0.05, 0.1, and 0.15 by firing at 1000 o C. Potassium is studied as a potential stabilizer due to large A - O bond length and its preference for prismatic co - ordination. Solid - state synthetic method was utilized to prepare potassium substituted compositions, [Na 0.5 - x K x ][Li 0.17 Ti 0.83 ]O 2 (x=0.05, 0.10, and 0.15). However, it was found that potassium does not show solubility in the studied material and tunnel - type impurity phase, with K 0.8 [Li 0.13 Ti 0.87 ]O 4 prototype structure, (ICSD reference: 89 - 5420) formed under the synthetic conditions as shown in Figure 3 . 23 b. Attempts to quench the powders rapidly from 1000 o C to room temperature to possibly freeze the meta - stable low - sodium content phases yielded identical impurity phases. It is likel y that the formation of the sodium deficient structures with large interlayer spacing is unfavorable, especially with solid - state reaction conditions involving high thermal energy. 3.8 Summary and future work In this chapter, titanate - based layer - structured quaternary oxide, sodium nickel titanate, is investigated as a novel sodium intercalation electrode. The solid - state synthetic methods were optimized using micron - sized and nano - sized precursors to prepare phase - pure powders with different particle morphol ogies. We studied the electrochemical properties of P2 - NT phase and - - 112 al cost benefits in large - scale battery manufacturing and reduce device cost. Electrochemical testing - 1/3 with reversible sodium intercalation/de - intercalation reactions in a solid - solution regime with a high redox potential of 3.7 - 1 provides a low redox potential of 0.7 V. The rate performance in the anodic region is better than the cathodic region due to higher amount of titanium in the tr ansition metal layer. The cathodic P2 - NT phase yields good capacity retention while the anodic P2 - NT phase has a capacity fade during cycling. This continuous capacity loss is attributed to catalytic breakdown of electrolyte components at low potentials an d requires further spectroscopic characterization to understand degradation mechanisms. In the cathodic regime, we have shown that manganese incorporation has a positive effect on the transport and electrochemical properties. We have demonstrated improved capacity retention at high currents, and achieved good cycleability. These performance gains are attributed to the following mechanisms: faster transport of charge - carriers (electron/holes and Na - ions) leading to higher electronic and ionic conduction, an d accelerated interfacial reaction kinetics as shown by EIS, PITT and DC measurements. P2 - NT and P2 - NMT phases have a dominant sodium transport mechanism and ionic conductivity is 5 orders of magnitude higher than electronic conductivity. Further improveme nts need to target achieving metallic - type conductivity by elemental substitution. EIS illustrated that interfacial resistance has a strong dependence on electrode potential and decreases by several orders of magnitude at low sodium content due to the form ation of more Ni 4+ states. We report a high sodium diffusion co - efficient in the range of 10 - 12 - 10 - 13 cm 2 /s. Impedance - coupled - cycling study showed that P2 - NMT has excellent capacity retention due to fairly stable charge transfer resistance, surface film r esistance and film capacitances. 113 Attempts to improve the stability of P2 - NT phases using calcium substitution (20%) were unsuccessful due to the formation of stable perovskite - structured calcium impurity phases during material synthesis. We did not observe solid evidence of solid solution formation with calcium even at lower substitutional level of 5%. It was identified that the anodic capacity can be improved substantially by utilizing sodium deficient compositions (Na x [Ni 1/3 Ti 2/3 ]O 2 where x<2/3). T hese highly non - stoichiometric phases were found to be unstable and could not be synthesized under traditional solid - state routes. Efforts to stabilize these phases that contain high vacancy concentration by incorporating larger potassium metal in the alka li layer results in phase separation with tunnel - type impurity oxides. Soft chemistry routes are often utilized to prepare metastable phases and it is likely that these potassium substituted phases can be synthesized under milder conditions. 39 114 BIBLIOGRAPHY 115 BIBLIOGRAPHY 1. Delmas, C., Fouassier, C. & Hagenmuller, P. Structural classification and properties of the layered oxides. Physica 99B, 81 85 (1980). 2. Johnson, C. S. et al. Structural Characterization of Layered LixNi0.5Mn0.5O2(00.7) Single Crystals. J. Supercond. Nov. Magn. 27, 1235 1238 (2014). 159 5. SYNTHESIS, STRUCTURAL AND ELECTROCHEMICAL CHARACTERIZATION OF Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 , AN O - TYPE LAYER - STRUCTURED MATERIAL 5.1 Introduction 5.1.1 O3 - type materials The electrochemical and structural properties of P2 - type titanate materials were investigated in the previous chapters. This chapter delves into characterizing the sodium intercalation/de - intercalation properties of O3 - type titanate composition, Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 . Like P2 - type materials, O3 - type materials also possess layered structure with alternating stacking of sodium and transition metal sheets . O3 - type materials have the prototypical - NaFeO 2 structure , iden tical to the well - studied and commercialized LiCoO 2 chemistry . 1 Quaternary O3 - type oxides are attractive sodium intercalation electrode candidates due to High theoretical cap acity due to higher sodium content i. High concentration of divalent transition metal s (compared to P2 composition) enhancing the electron hopping and consequently electrical conductivity ii. Good ion transport properties facilitated by the 2D sodium slabs Howeve r, the ion con ductivity of O3 - type are lower than the P2 - type materials due to the sodium octahedral co - ordination and lower vacancy concentration. This has been established using c omputational and experimental studies. 2 But, solid - state ion tra nsport is often not the rate limiting step for titanate - based layered materials . The higher divalent transition metal concentration in the O3 materials promotes the formation of a percolating network and possibly ease limitations due to charge carrier tran sport processes . 160 5.1.2 Literature O3 - type materials are known to undergo multiple phase transitions , sodium/vacancy ordering during various stages of sodium deintercalation and these processes often lead to poor reversibility, and capacity retention . - NaFeO 2 has an initial reversible capacity of 80 mAh/g but the capacity drops by 25% in 30 cycles with a relatively low charging voltage of 3.4 V. 1 Extending the charging voltage to 4.0 V worsens the c apacity fade to 80% at the end of 10 cycles . This large irreversible capacity loss has been attributed to the migration of iron to the sodium layers and subsequently blocking sodium transport . It is interesting to note that this m igration is sue is energetically favorable for O3 materials and unfavorable for P2 materials as the former structures retain octahedral co - ordination for both transition metal and sodium sites. O3 - NaCoO 2 and O3 - Na[Fe 0.5 Co 0.5 ]O 2 shows a capacity decay of 7% in 30 cycles, and 16% in 50 cycles, respectively. 3 O3 - Na[Ni 1/3 Co 1/3 Fe 1/3 ]O 2 has 5% capacity drop in 2 0 cycles when cycled till 4.2 V with good electrochemical reversibility . The presence of large amount of cobalt promotes charge carrier mobility and hence this composition offers excellent high power performance . Further long - t ime cycling is required for this composition 4 In 50 cycles, the capacity fade in O3 - Na[Ni 0.5 Mn 0.5 ]O 2 was reported to be 6 0 % and it was further improved to 3 4 % by electrolyte optimization by the addition of F luorinated Ethylene C arbonate (FEC) additive . 5 , 6 FEC additive effectively suppresses fade mechanisms by arresting side reactions as indicated by improved coulombic efficiency . Nevertheless, even w ith FEC additive, the in trinsic fade mechanism is still quite substantial. O3 - Na[Ni 1/3 Mn 1/3 Co 1/3 ]O 2 cathodes (NMC) provide excellent capacity retention with no noticeable fade when cycled between 2 - 3.75 V for 50 cycles at C/5 rate. 7 1 61 From these literature reports, it becomes apparent that capacity fade is a common issue for many O3 - type cathode materials and addressing this is crucial for further advancement of high - voltage, high - capacity cathode materials for Na - ion batteries. In addition to the aforementioned materials, several quaternary compositions were reported in light of growing interest in Na - ion battery chemistry. 8 11 5.1.3 Research Strategy The focus of this research work is to investigate the i nfluence of titanium on the elect rochemical properties and phase transitions in Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 composition using electroanalytical techniques suc h as impedance testing and step - potential technique, and ex - situ XRD testing . We anticipate that t he presence of tetravalent titanium ato ms provides a stabilizing influence in the transition metal layer due to the increased covalent bonding character (in Ti - O bonds) resulting in better cycleability than other O 3 - type transition metal oxides. R ecently, a titanate - based O3 material, Na[Ni 0.5 Ti 0.5 ]O 2 , was reported to have excellent capacity retention (only 6.8% drop in 100 cycles with 4.0 V charging voltage), high coulombic efficiency (96%), and good high - rate performance (90.5 mAh/g at 1C). However, this study lacks structural characterizatio n to identify the presence and onset of phase transitions that are directly correlated with the reversibility of the sodium intercalation reactions and cycle - life performance . Also, the powders prepared in this study contained NiO impurity phase ( electroch emically inert) that could be eliminated by optimized synthetic methods to achieve improved reversible capacitites . 12 162 5.2 Material synthesis & electrode f abrication The O3 NT oxide powders wer e prepared by firing the SPEX dry milled precursors at 1000 o C in a box furnace for 10 hours duration. The XRD pat tern confirms that the synthesized powders are phase - pure (matches with the PDF reference code : 01 - 070 - 6690 ) without any impurity oxide phases as illustrated in Figure 5 . 1 a. The reproduction of experimental XRD pattern by the predicted Bragg reflections confirms that this ma terial crystallizes with space group R - 3m. T he calculated lattice parameters are: a=2.9983 Å , and c=16.1831 Å . The refined sodium occupancy value of 0.86 is fairly close to the target composition of 0.9. The powders were comprised of agglomerates of micro n - sized particles as shown by the SEM inset micrographs. The individual particles appear fairly dense and are micron - sized . The film electrodes consisted of the oxide particles embedded in a highly c onductive network formed by carbon particles as depicted in the SEM and EDS mapped images ( Figure 5 . 1 b and c ). The titanium is uniformly distributed in the individual particles demonstrating the absence of secondary titaniu m - rich phase / - depleted phases. Composite electrode films were fabricated consisting of 80/10/10 weight ratio of active material, conductive carbon (Timcal C65) and Polyvinylidene fluoride binder . N - M ethyl - 2 - pyr r olidone (NMP) was used as solvent for casting the film electrodes using the conventional doctor - blade technique. Aluminum was utilized as the substrate for casting the electrode films. After casting, the electrode films were dried using IR lamp overnight to remove the NMP with the same protocol as discussed in previous chapters . After overnight drying, the film electrodes were 2 O and O 2 ) for storage before subsequent electrochemical testing. 163 (a) (b ) (c ) Figure 5 . 1 (a) Rietveld refinement of O3NT powders prepared by solid - state reaction at 1000 o C for 10 hours firing duration with inset figure showing powder morphology, (b) SEM morphology of composite film electrode showing O3NT particles embedded in a carbon network formed by conductive additive, and (c) EDS mapping of powders showing elemental distribution of carbon and titanium; refinement was performed with lab powder diffraction dataset u sing TOPAS software by Bruker Corporation . 13 16 164 5.3 Electrochemistry 5.3.1 V oltage - composition curves : Galvanostatic Testing O3 NT has a high theoretical capacity of 229 mAh/g for complete removal of 0.9 moles of sodium per mole of the active material . Ideally, the sodium de - intercalation/intercalation reactions would follow a solid solution mechanism over a broad range of sodium content with the activation of Ni 2+/4+ redox couple as depi cted in Equatio n 5 . 1 . 5 . 1 ( a ) ( b ) Figure 5 . 2 Voltage - composition curve of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 under galvanostatic 1 st charging cycle is due to side reactions. Th e charge/discharge curves of O3 NT material tested under galvanostatic conditions of C/50 is shown in Figure 5 . 2 using the voltage - composition curves . The electrolyte utilized for this testing consisted of 0.5 M NaPF 6 salt dissolved in Ethylene Carbonate (EC)/Diethyl Carbonate (DEC) 165 (50/50 vol%) . A nodized aluminum and copper rods function as current collectors for working ele ctrode and metallic sodium, respectively. When the charging voltage is limited to 4.2 V, the 1 st charg e and discharg e capacities are 222.7 mAh/g, and 99.7 mAh/g, respectively. The poor faradaic efficiency in the 1 st cycle (44.8%) is due to electrolyte side reactions possibly facilitated by the active material . The 5 th cycle charge and discharge capacities are 127.5 mAh/g, and 108. 4 mAh/g, respectively , with improved faradaic efficiency of 85% . The voltage profiles suggest that sodium intercalation/de - intercalation reactions can be sustained reversibly in the evaluated voltage window. Increasing the charging cutoff voltage to 4.5 V yields 1 st charge and discharge capacities of 264 mAh/g, and 112.6 mAh/g, respectively. The first c harging capacity is clearly higher than the theoretical capacity due to additional side reaction contributions . The discharge capacity improves slightly when the charging voltage is increased to 4.5 V due to the formation of more vacant sites in the chargi ng step. However, the faradaic efficiency becomes slightly worse ( 1 st cycle: 42.7%) in the highly oxidative potenti al regime. The voltage curves have small plateaus at 3.1 and 4.2 V, respectively, and a sloping profile in the remaining regions. The voltage plateaus and sloping regions are associated with two - phase and solid solution reactions , respectively, and are discussed further in the ex - situ XRD section. 5.3.2 Phase trans itions: Ex - situ XRD Experimental methods for ex - situ XRD characterization discussed in Chapter 3 were utilized for O3 NT electrodes as well . The film electrodes were tested w hile exposed to ambient conditions for the preliminary trials. O3 - type materials undergo phase transitions when sodium is extracted from the pristine composition and the crystal structure of the octahedral starting phase (O3) and prismatic low - sodium content phase (P3) are depicted in Figure 5 . 3 . Due to the difference in the 166 oxygen stacking sequence, the O3 and P3 phases have di fferent sodium co - ordination environment s . (a) (b) Figure 5 . 3 Crystal structure of (a) O3 phase with octahedral co - ordination of sodium, and (b) P3 phase with prismatic co - ordination of sodium demonstrating the different oxygen stacking sequence. The initial O3NT phase with space gro up R - .9 curve in Figure 5 . 4 . Removal o f small amount s of sodium (10%) c auses the 16.5 o peak intensity, characteristic of O3NT pristine phase, to drop considerably while a new peak forms at 15.8 o . This peak splitting is associated with phase transformation from the initial O3 phase to P3 phase. This contrasts with other O3 compositions that ha ve the phase transition sequence of O3 (hexagonal) - (monoclinic) - P3 (hexagonal). 17 At the measured compositions, there is no evidence of the formation of the low - . The measured X RD reflections can be matched well with a two - phase mixture of O3+P3 as shown in Figure 5 . 5 a . It is likely that the O3 - occurs in much smaller composition region of . Continued sodium extraction 167 results in higher amount s of P3 phase at the expense of O3 phase as indicated by decreasing intensities of 16.5 o XRD peak . Figure 5 . 4 Ex - to the unidentified phase formed due to moisture intercalation of the low - sodium content phases. The O3 - P3 phase change involves gliding motion of oxygen layers to compe nsate the repulsive interactions between the closed - pack oxygen layers at low sodium contents. Hence, P3 phases are expected to have a higher Na - O bond length than the O3 phase. This is confirmed by longer average Na - O bond length in P3 phase (2.582 Å ) tha n O3 phase (2.451 Å ). At s odium content (x) of 0.3 9 , pure P3 phase is obtained and further sodium removal essentially leads to a solid - solution zone in the composition region of and the refined patterns are shown in Figure 5 . 5 b. When charging voltage is increased beyond 4.2 V, O3NT electrodes show 168 a splitting of the 15.6 o peak and the resulting new peak has significant broadening. This peak splitting is not consistent with any P - rela ted phases that were reported for other O3 - materials. We suspect that this broadening is due to atmospheric interaction of the low sodium content P - phases (in composition zone of ) that form at high ly oxidative potentials . These phases can potentially accommodate water in the structure due to the large available spacing in the alkali slabs . Such water intercalation processes can result in lo ss of crystallinity due to exfoliation and this is consistent with the br oad ening of the XRD patterns . Efforts to mitigate moisture interaction by transferring the powders in inert atmosphere and limiting the atmospheric exposure to only a short time window (~ 10 minutes) with fast - scans we re unsuccessful due to the extremely high moisture sensitivity of these phases . Future work to obtain structural information on the low - sodium cont ent phases would require designing a hermetic in - situ electrochemical setup with XRD testing capability . 169 (a) (b) Figure 5 . 5 Rietveld refinement of O3NT electrode (Na x [Ni 0,45 Ti 0.55 ]O 2 ) prepared by using TOPAS software from Bruker Corporation , insert figures depict sodium co - ordination environment and Na - O bond lengths. 170 5.3.3 Equilibrium voltage profiles and solid - state diffusi vity estimation : Potentiostatic Intermittent Titration Technique (PITT) The PITT curves of O3 NT material are d epicted in Figure 5 . 6 a and b . Overall, the voltage profile trends are consistent with the galvanostatic testing and show a smaller overpotential due to the low limitin g current value (C/100). After the 1 st cycle, the 4.2 V charge cutoff experiment has a very low voltage hysteresis between charge and discharge curves. However, the 4.5 charge cutoff experiment exhibits marked voltage hysteresis due to the insufficient rev ersibility of the low sodium content phases. The fairly good overlapping of the voltage profiles for both experiments, after the initial cycle, suggest that the O3NT material does not undergo deleterious structural changes due to repeated charge/discharge. The electrochemical stability for longer durations is further evaluated at faster charging rates in the cycling study. The measured current transients while imposing the step potential in the PITT study are utilized for evaluating the solid - state diffus ion co - efficient of sodium ions. Typically, an intercalation material undergoing phase transformation has two distinct current profiles and the rate of movement of the phase boundary essentially limits the diffusion process. 18 However, the experimental current decays, shown in Figure 5 . 6 c and d, do not match a phase separation material and is more consistent with an intercalation - type material. This is likely due to rapid movement of the phase boundary separating the O3 and P 3 phases . We calculate the apparent chemical diffusion co - efficient using the theory discussed in chapter 3 for O3NT phase at various sodium contents. The fairly linear current regions further validate the applicability of this model. 171 (a) (b) (c) (d) Figure 5 . 6 (a) Voltage profile of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 using Potentiostatic Intermittent Titration Technique (PITT) with C/100 current and 4.2 V voltage limit, (b) Voltage profile using PITT with C/100 current and 4.5 V voltage limit, (c) current transient linear fitting with 4.2 V limit, and (d) current tran sient linear fitting with 4.5 V limit. 172 Figure 5 . 7 Solid - state diffusion co - efficient of sodium ions in O3NT material using potentiostatic intermittent titration testing. The calculated sodium diffusion co - efficients are in the order of 5x 10 - 11 cm 2 /s, considered reasonably fast for solid - state transport in intercalation materials as shown in Figure 5 . 7 . The increase in diffusivity with voltage is consistent with the mechanism of faster ion hopping due to the creation of extra sodium vacant sites. The solid - solution model does not apply at high v oltages as evidenced by non - linearity in current transient in Figure 5 . 6 d. 5.3.4 Cycling study 5.3.4.1 Capacity retention at different voltage cutoffs O3NT material was cycled bet ween 2. 0 - 4.0 V at C/10 rate for initial 10 cycles and C/2 rat e subsequently as shown in Figure 5 . 8 a. It is apparent that even with a relatively low voltage cutoff of 4.0 V, there is some irreversible side reactions in the initial cycles but the faradaic efficiency improves to 99% in the subsequent cycles at C/2 rate demonstrating excellent reversibility. The 173 discharge capacity values were measured to be 58.7 mAh/g, and 55.2 mAh/g at 25 th and 11 5 th cycle, respectively, corresponding to capacity decay o f 6%. For the electrode cycled to 4.2 V, the faradaic efficiency remains at 54% in the 1 st cycle showing that side reactions did not increase between 4.0 and 4.2 V experime nts ( Figure 5 . 8 b) . This results in a higher reversible capacity of 75.1 mAh/g, and 74. 2 mAh/g at 25 th and 11 5 th cycle, respectively, with capacity retention of 99 %. Typically, the capacity fade is directly correlated with the amount of extracted sodium ions due to stra in induced failure mechanisms. However, t he cycling t rend s do not match this theory and this discrepancy is likely due unaccounted random errors. Nev ertheless, we show that the titanate - based O3NT material has excellent capacity retention. At charging voltage s beyond 4.2 V , capacity values are still stable, contrary to other O3 - materials, as shown in Figure 5 . 8 c. But the practical discharge ca pacit ies are fairly r estricted at these high operating voltages due to resistive contributions p ossibly from passivation films. This improved capacity retention is attributed to the stabilizing influence of titanium in the transition metal layer for this O3NT composition. 174 ( a ) ( b ) (c) ( d ) Figure 5 . 8 Capacity study of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 material: (a) 4.0 V charging limit, (b) 4.2 V charging limit, (c) 4.4 V charging limit, and (d) Variation of charging capacity with different cut - off voltage limit. The charging capacit i es are plotted as a function of cut - off voltage in Figure 5 . 8 d. T he 1 st cycle charging capacities increase with charging cutoffs, as expected, with the maximum at 4.5 V c harging voltage. However, 10 th , 25 th and 55 th cycle charging capacity curves have a maximum at 4.2 V and the electrodes charged beyond 4.2 V have lower capacity . This suggests that the resistive 175 contributions possibly due to passivation films limit the capacity values at the high charging voltages (>4 .2 V) . Reg ardless of the passivation film effects , the presence of titanium in the O3NT material arrests the capacity fade significantly. 5.3.4.2 Impedance spectra evolution during cycling The impedance spectra of the O3NT material collected after chargin g the films to 4.2 V during the cycling study are depicted in Figure 5 . 9 a . The Nyquist plots have the following features: two over - lapping arcs in the high frequency region , one arc in the medium frequency region and diffusional tail in the low frequency region . The Bode plot in Figure 5 . 9 b provides the total impedance over the complete frequency domain and the arcs correspond to plateaus. The two high - f requency arcs are attributed to primary and secondary surface film resistan ce coupled with film capacitance in parallel RC circuit arrangement. The medi um frequency arc is related to interfacial charge transfer reactions coupled with double layer capacitance . This equivalent circuit fitting approach is similar to P2NT phase imped ance analysis discussed i n Chapter 3 . The circuit model, depicted in the inset Figure 5 . 9 c, describes the physicochemical processes . Though Warburg component would ade quately describe the linear tail at the low frequencies, it is not considered for diffusional analysis as the tail gets overshadowed during electrode cycling. Repeated electrode cycling does not affect the interfacial kinetics and primary surface film r esistances as shown in Figure 5 . 9 c. This is consistent with the invariant high frequency arc and high frequency plateau in the Nyquist and Bode plots. However, the sec ondary surface film resistance increases gradually during cycling and this trend matches with the growing medium frequency arc that eventually overshadows the diffusional tail for a cycled electrode (n>80). The double layer capacitance (Cct), and secondary surface film capacitances (Csf2) are measured in mF range and the primary surface film capacitance (Csf1) fluctuates in mF - µ F range ( Figure 5 . 9 d) . 176 The latter fluctuation might be related to film instability. The demonstrated excellent c apacity retention of O3NT phase is attributed to relatively minor change s in the physicochemical pa rameters that govern the sodium intercalation processes from impedance spectroscopy. The impedance parameters also show that interfacial charge transfer process is the slow rate - limiting processes for O3NT phase as it provides the largest resistive contrib ution. Quaternary compositions with cobalt, manganese substitutions can further promote the rate of charge carrier transport and further reduce the interfacial charge transfer resistance. 177 (a) (b) (c) (d) Figure 5 . 9 (a) Nyquist plot of O3NT electrodes charged to 4.2 V during cycling, (b) Bode plots of O3NT electrodes charged to 4.2 V during cycling, (c) Variation of electrolyte resistance (R e ), surface film resistances (Rsf1 and Rsf2), an d charge transfer resistance (Rct) during cycling, and (d) Variation of double layer capacitance (Cdl) and surface film capacitances (Csf1 and Csf2) . 178 5.3.4.3 Sodium metal de activation: Prolonged cycling Figure 5 . 10 represents the variation of the maximum counter electrode potential during each cycle with cycle number during the cycling study of O3NT electrode with 4.2 V cutoff . The counter electrode potential is meas ured with respect to the sodium reference electrode . Under open circuit conditions , the counter electrode potential would be the same as the reference electrode and the measured V would read zero. At galvanostatic conditions, the voltage profile shows a s mall overpotential caused by kinetic or mass transport limitations at the counter electrode/electrolyte interface . Figure 5 . 10 Variation of peak counter electrode potential with cycle number during galvanostatic cycling measured against Na/Na + reference electrode; ideally the measured voltage would be zero at the counter electrode as it has the same chemical potential as the refe rence electrode. Upon repeated cycling, it is evident that the counter electrode peak voltage shifts drastically from zero towards 3 V. They measured voltage values are fairly close to the standard reduction potential of copper that is used as the curren t collector rod. During a discharge run, oxidation reaction needs to be supported at the counter electrode and this is typically through oxidation of 179 metallic sodium to sodium ions. When the sodium ions gets plated (reduction) during the subsequent chargin g run, the deposited sodium tends to be fairly powdery and has poor adhesion resulting in loss of active sodium metal. Repeated cycling leads to continual depletion of the active sodium metal surface due to this de - activation process. Further cycling leads to copper rods getting etched and large overpotential (typically beyond cycle number~120). This effect manifests as a capacity fade during the cycling studies when in actuality the issue is only due the counter electrode. The continual deactivation of the counter electrode has limited the cycling runs to ~100 cycles. One proposed method to avoid this is to directly use an intercalation - based anode and av oid sodium plating completely. However, this requires careful selection of a suitable anode intercalatio n material with good electrochemical stability and little irreversible capacity losses. 5.3.5 Rate testing O3 NT material provides discharge capacity values of 74.9 mAh/g, 71.2 mAh/g, 60 mAh/g, 47.8 mAh/g, and 30.4 mAh/g at C/10, C/5, C2, 1C, and 2C , respectivel y as shown in Figure 5 . 11 a. The reduced capacities at high cur rents is linked to increasingly steeper sloping volta ge curves ( Figure 5 . 11 b) at relatively long charge/discharge timescales. This is consistent with the impedance measurements that demonstrate interfacial charge transfer limitations at medium and low frequency region s . Nevertheless, O3NT phase has improved capacity values at higher currents than P2 - NT phase due to the higher Ni/Ti ratio in the latter. The charge carrier transport (electron/holes) is facilitated primarily through the nickel atoms (Ni 2+ / 3+ states) in the transition metal layer. Hence, a higher concentration of the nickel atom s results in the formation of a percolating network promoting facile movement of the charge carrier species. 180 (a) (b) Figure 5 . 11 Rate performance testing of O3 - Na 0.9 [Ni 0.45 Ti 0.55 ]O 2 material with 4.2 V cut - off voltage (a) capacity retention, and (b) voltage profiles. 5.4 Summary and future work In this chapter, we study the electrochemistry of layered structured materials with octahed ral co - ordination (O3) of sodium as they possess high theoretical capacity. These O3 materials suffer from capacity fade due to electrochemically driven phase separation and we characterize titanium based composition (O3NT) to impart structural stability b y utilizing the covalent bonding character of titanium. Phase - pure O3NT powders were prepared by solid - state reaction and characterized by XRD and SEM. Galvanostatic testing confirms that this composition supports reversible sodium intercalation/de - intercalation by activating the Ni 2+/4+ redox coup le while titanium remains invariant at 4+ valence state in the transition metal layer. PITT shows that a small voltage hysteresis between charge/discharge curves when the material is cycled with 4.2 V cutoff validating the electrochemical reversibility. The hysteresis becomes larger for films charged to 4.5 V due to irreversibility associated with newly formed phases. Ex - situ XRD shows that the O3 181 phase converts to P3 phase for small sodium removal resulting in the formation of a two - phase region and this region occurs for . This phase transformation involves gliding motion of the closed - pack oxygen layer s due counteract the increasing repulsive interactions between oxygen layers . solid - solution domain is reported for . The quality of XRD pattern on powders with even lower sodium content were insufficient for structural analysis due to moisture interaction effects. Cycling studies support that operating in the O3 - P3 domain provides excellent capacity retention of 99% at the end of 115 charge/discharge cycles at C/2 rate. This confirms that titanium does provide a strong stabilizing influence by suppressing the capacity fade. The pre sence of P3 solid - solution regime over a broad voltage window allows for structural stability without significant capacity losses. Electrodes charged to high voltage of 4.5 V also have stable albeit lower capacity values. Impedance spectroscopy confirms th at the rate of change of the physicochemical parameters that govern the sodium intercalation processes is relatively small supporting the excellent capacity retention characteristics. Rate study confirms that O3NT phase has improved performance compared to the P2NT phase due to the presence of higher nickel content in the transition metal layer aiding the charge carrier transport processes. Structural characterization of the low - sodium content phases is essential to identify the phases and this requires fu rther in - situ electrochemical - XRD testing. Given that the titanate - based composition provide s stable capacity retention at high voltages (4.5 V) despite the phase transitions , future work can attempt at coating the oxide particles with an ion - conducting/in ert passivation layer that can inhibit side reactions by limiting the electrolyte contact area . Potential coating materials include sodium zirconium silicophosphate , sodium yttrium silicate, - alumina, 182 aluminum oxide, etc. 19 23 Alternatively, e lectrolyte additives such as FEC can also be employed to reduce the high - voltage side reactions. These approaches can potentially provide higher reversible capacities by allowing high voltage operation while utilizing the inherent stability effect provided by the titanate composition . 183 B IBLIOGRAPHY 184 BIBLIOGRAPHY 1. Yabuuchi, N., Yoshida, H. & Komaba, S. Crystal structures and electrode performance of alpha - NaFeO2 for rechargeable sodium batteries. Electrochemistry 80, 716 719 (2012). 2. Lee, D. H., Xu, J. & Meng, Y. S. An advanced cathode for Na - ion batteries with high rate and excellent structural stability. Phys. Chem. Chem. Phys. 15, 3304 12 (2013). 3. Yoshida, H., Yabuuchi, N. & Komaba, S. NaFe0 .5Co0.5O2 as high energy and power positive electrode for Na - ion batteries. Electrochem. commun. 34, 60 63 (2013). 4. Vassilaras, P., Toumar, A. J. & Ceder, G. Electrochemical properties of NaNi1/3Co1/3Fe1/3O2 as a cathode material for Na - ion batteries. El ectrochem. commun. 38, 79 81 (2014). 5. Komaba, S. et al. Study on the Reversible Electrode Reaction of Na1 - xNi0.5Mn0.5O2 for a Rechargeable Sodium - Ion Battery. Inorg. Chem. 51, 6211 20 (2012). 6. Komaba, S. et al. Fluorinated Ethylene Carbonate as Electro lyte Additive for Rechargeable Na Batteries. ACS Appl. Mater. Interfaces 3, 4165 4168 (2011). 7. Sathiya, M., Hemalatha, K., Ramesha, K., Tarascon, J. - M. & Prakash, A. S. Synthesis, structure, and electrochemical properties of the layered sodium insertion cathode: NaNi1/3Mn1/3Co1/3O2. Chem. Mater. 24, 1846 1853 (2012). 8. Huili, P., And, H. Y. - S. & Liquan, C. Room - Temperature Stationary Sodium - Ion Batteries for Large - Scale Electric Energy Storage. Energy Environ. Sci. (2013). doi:10.1039/C3EE40847G 9. Slater, M. D., Kim, D., Lee, E. & Johnson, C. S. Sodium - Ion Batteries. Adv. Funct. Mater. 23, 947 958 (2013). 10. Palomares, V. et al. Na - ion batteries, recent advances and present challenges to become low cost energy storage systems. Energy Environ. Sci. 5, 5884 (2012). 11. Ellis, B. L. & Nazar, L. F. Sodium and sodium - ion energy storage batteries. Curr. Opin. Solid State Mater. Sci. 16, 168 177 (2012). 12. Yu, H., Guo, S., Zhu, Y., Ishida, M. & Zhou, H. Novel titanium - based O3 - type NaTi(0.5)Ni(0.5)O2 as a cathode material for sodium ion batteries. Chem. Commun. 50, 457 459 (2014). 13. TOPAS V4: General Profile and structure analysis software for powder diffracti on data. - . (2008). 185 14. Cheary, R. W. & Coelho, A. A. A fundamental parameters approach to X - ray line - profile fitting. J. Appl. Cryst. 25, 109 121 (1992). 15. Cheary, R. W., Coelho, A. A. & Cline, J. P. Fundamen tal parameters line profile fitting in laboratory diffractometers. J. Res. Natl. Inst. Stand. Technol. 109, 1 25 (2004). 16. Kern, A., Coelho, A. A. & Cheary, R. W. Convolution based profile fitting - Diffraction analysis of the microstructure of materials . 17 50 (2004). 17. Physical and Electrochemical Studies. J. Solid State Chem. 117, 111 117 (1996). 18. Lai, W. & Ciucci, F. Thermodynamics and kinetics of phase transforma tion in intercalation battery electrodes - phenomenological modeling. Electrochim. Acta 531 542 (2010). doi:10.1016/j.electacta.2010.09.015 19. Clearfield, A., Jerus, P. & Cotman, R. N. Hydrothermal and solid state synthesis of sodium zirconium silicophosph ate. Solid State Ionics 5, 301 304 (1981). 20. Shannon, R. D., Taylor, B. E., Gier, T. E., Chen, H. - Y. & Berzins,, T. Ionic Conductivity in Na5YSi4012 - Type Silicates. Inorg. Chem. 17, 958 964 (1978). 21. Whittingham, M. S. & Huggins, R. A. Measurement of S odium Ion Transport in Beta Alumina Using Reversible Solid Electrodes. J. Chem. Phys. 54, 414 (1971). 22. Cho, J., Kim, Y. J. & Park, B. Novel LiCoO2 Cathode Material with Al2O3 Coating for a Li Ion Cell. Chem. Mater. 3788 3791 (2000). 23. Scott, I. D. et al. Ultrathin Coatings on Nano - LiCoO2 for Li - Ion Vehicular Applications. Nano Lett. 11, 414 418 (2011). 186 6. CONCLUSIONS Designing low - cost, long cycle life battery devices remains a prime challenge for stationary and grid - sector applications. Na - ion batteries are actively researched to address the gaps in these market areas and is anticipated to provide a low - cost, sustaina ble solution as they are based on inexpensive, plentiful and geographically widespread sodium resources. Characterization of novel functional intercalation - type materials remains a research challenge for device development. Availability of multiple electro de material choices would diversify the product performance specifications and address a larger market segment with different performance metrics. In this research work, inorganic materials with various structure types such as tunnel - and layer - related pha ses were investigated as novel sodium intercalation electrodes. Using impedance spectroscopy, we show that perchlorate based electrolyte (1M NaClO 4 in EC/DMC) had a continual growth of surface films contributing to increasing film resistance at open circu it conditions. Fluorine based electrolyte (0.5M NaPF 6 in EC/DEC), on the other hand, had a lower film resistance with a gradual growth rate of the surface films demonstrating better passivation stability. Cyclic voltammetric testing showed that anodized al uminum (with exposed aluminum cross - section to contact film) provided excellent corrosion resistance as indicated by lower oxidative peak currents than nickel, stainless steel, graphite, titanium, and plain aluminum. Electrochemical methods were validated using well - studied Na 0.7 CoO 2 and LiFePO 4 electrode materials using in - lab synthesized materials. The electrochemistry of novel, tunnel - type materials: sodium nickel phosphate (SNP) and sodium cobalt titanate (SCT) were evaluated. Reversible sodium de - inter calation/intercalation reactions were shown in SNP using Ni 2+/4+ redox couple with average voltage of 2.0 V vs. Na/Na + with 50% of theoretical capacity (~50 mAh/g). A large voltage 187 hysteresis of 1V was obtained between charge/discharge curves. Ex - situ XRD testing confirmed that the material retained the starting structure due to the inherent structural stability of polyanionic material classes. PITT testing of SCT shows sloping, symmetric charge/discharge curves between 3.6 - 4.2 V indicative of intercalation - type reactions with good faradaic efficiency of 86%. The voltage hysteresis is 1.4 V in fluorine based electrolyte when tested at 70 o C and increases to 1.6 V in perchlorate based electrolyte when tested at room temperature. The voltage hysteresis is repo rted to be inherent to the material chemistry and a thermally activated process. We conclude that the slow solid state diffusion of sodium ions in SNP and SCT causes the large voltage polarization and further improvements would require significant particle size reduction to reduce the diffusional length scale. Layer - structured titanate materials with prismatic (P2) and octahedral co - ordination (O3) of sodium atoms were investigated as potential sodium intercalation electrodes. P2 - NT and P2 - NMT materials we re prepared by optimal solid state reaction techniques to obtain reasonably phase pure powders. High cathodic redox potential of 3.7 V vs. Na/Na + for Ni 2+/4+ and low redox potential of 0.7 V vs. Na/Na + for Ti 4+/3+ is reported for P2 - NT phase demonstrating its bifunctional nature. Using P2 - large - scale manufacturing. The sloping voltage profile between 2 - 4.2 V, and ex - situ XRD testing confirms that the P2 - NT phase retains the initial structure without any phase change. Good reversibility is demonstrated in P2 - NMT phase as well with few additional peaks at E>4.0 V possibly due to sodium ordering processes. AC and DC testing on sintered pellets show a wide separation between the electronic and ionic conductivity curves with the former being ~4 orders faster. Impedance spectroscopy of film electrodes show a decreasing interfacial resistance with charging voltage from equivalent circuit modeling. P2 - NMT electrodes were shown to poss ess 188 faster interfacial reaction kinetics than P2 - NT electrodes over the complete voltage range. PITT and EIS data supports relatively fast solid state diffusion of sodium ions (10 - 12 - 10 - 13 cm 2 /s), as expected for the prismatically co - ordinated layered oxid es. P2 - NMT electrodes were found to provide significantly improved capacity retention than P2 - NT electrodes at higher currents due to the faster interfacial kinetics and enhanced charge - carrier transport processes. These materials are shown to possess good capacity retention in the cathodic regime. However, P2 - NT phase has a pronounced capacity fade in the anodic regime and this is likely associated with electrocatalytic reduction induced by nickel species. This can potentially be avoided by forming a prote ctive surface film to limit electrolyte exposure. Attempts to incorporate calcium and potassium in the P2 structure by traditional solid - state synthetic routes to stabilize the structure for improved performance were unsuccessful due to the formation of pe rovskite - and framework - related impurity phases. The structural properties of P2 - NT material was evaluated using experimental neutron diffraction and computational studies. Rietveld refinement provided average structural information while computational methods based on atomistic simulations provid ed both average and local structural details. The Bragg reflections of space group P63/mmc (# 194) reproduced the experimental patterns fairly accurately and utilized as a good structural model. The measured anisotropy in thermal expansion co - efficient val ues: = 1.83 ×10 - 5 Å /K, =17.65 ×10 - 5 Å /K was reported to be due to the layered configuration with weaker through - plane bonding. Na f sites were found to contain lower occupancy than Na e sites due to higher repulsive interaction with the in - line tr ansition metal atoms. The Na f site occupancy was found to gradually increase (0.1699 0.1751) as the testing temperature is raised from 15 300K. The higher thermal contribution at ambient conditions was found to make the unfavorable sodium sites slightly mo re 189 accessible. The experimental sodium nuclear densities displayed considerable oblateness along the ab - plane while oxygen nuclear densities remained fairly spherical. Buckingham and Morse - type interatomic potential models provided fairly accurate structur al representation of various binary and ternary oxides, and the target phase. Energetics study confirmed that equally distributed structures had the lowest energy for both potential models. The simulated nuclear densities matched the experimental patterns fairly well. The simulated patterns were also found to contain additional density spots due to displacive movements and this effect is attributed to the formation of stacking faults. The simulated and pseudo - experimental PDF patterns showed some difference s due to the averaging effect in Rietveld refinement and further analysis would require experimental PDF measurements. These potential models would act as a good framework for further model improvement using Ab - initio energy hypersurface fitting approaches and MD studies to investigate the dynamics of sodium transport. The electrochemical and structural properties of O3NT material was characterized with the potential of realizing higher reversible capacities than P2 materials. The stabilizing influence of titanium to reduce the capacity fade was evaluated. Galvanostatic testing and PITT confirm the excellent electrochemical reversibility with small voltage hysteresis with charging voltage of 4.2 V. However, higher voltages were found to lead to affect the r eaction reversibility. Ex - situ XRD testing show that the O3 phase transforms to P3 phase in composition range of . . Further extraction of sodium results in the formation of an unknown phase possibly with high moisture sensitivity. We show that cycling the electrode in the O3 - P3 domain provides for an excellent capacity retention of 99% in 115 cycles at C/2 rate. Impedance spectroscopy confirms that the physic ochemical parameters such as surface film resistance and interfacial charge transfer 190 remains fairly invariant during cycling contributing to the excellent capacity retention. We also report that O3NT phase has good capacity retention when cycled in higher charging voltages (>4.2 V) and these enhanced capacity retention is attributed to stronger bonding character of titanium. However, the capacity values become reduced in these high voltage cutoff studies possibly due to the formation of resistive surface fi lms by electrolyte breakdown at highly oxidative potentials. Future work involving masking the surface of the active particles can potentially inhibit these side reactions and provide improved capacity retention. Rate studies confirm that O3NT electrodes p rovide improved capacity retention than P2NT electrodes due to the presence of higher nickel providing a percolating network for fast transport of charge carrier species.