Glassy Oxides as Positive Electrode Materials for Sustainable and High Energy Density Li-Ion and Na-Ion Batteries: First Attempts to Establish Relationships between Glass Compositions, Physicochemical Characterizations and Electrochemical Properties with Machine Learning and Design of Experiments Approaches

Abstract

Today, rechargeable lithium-ion batteries (LIBs) represent one of the best alternatives to reduce our dependency on fossil fuels. LIBs are using cathode materials based on polycrystalline oxides or polyanionic compounds [1]. However, their performances can be limited by their crystalline structure when other compounds suffer from metallic dissolution and irreversible phase changes during cycling [2]. To overcome these key shortcomings, implementing glasses or glass-ceramics as cathode materials is an interesting approach [3]. Glasses have a structure composed of more free volume, and, in the literature that is mainly concerning glasses containing Vanadium transition metal, the glasses can so accept a large amount of lithium [4] and easily accommodate structural changes upon lithium ions extraction/insertion [5]. Furthermore, glass production is a scalable process and can be commercially easier to implement than most of synthesis processes of conventional cathode materials.The main objective targeted is to develop high capacity positive electrodes (up to 300mAh/g) for Li-ion batteries based on glassy oxides without critical transition metals such as Cobalt and Vanadium. The ambitious target is to exceed 1000Wh/kg at the active material level and so enable energy densities of 300-350 Wh/kg at the Li-Ion cell level. Additionally, the electrochemical behavior of some alkali-free glass compositions will be evaluated in sodium-ion configuration. The major scientific objective is to correlate the most relevant physicochemical properties of glass with the corresponding electrochemical performances using machine learning approaches or design of experiment methodology. For a better understanding of involved processes, three oxides glass families are studied as promising cathode materials for sustainable and high energy density lithium batteries. These oxides are composed of: i) a glass network-forming element (Si, B, P) leading to the 3 different families, ii) a transition metal (Mn, Fe, ...) and optionally iii) an alkali (Li, Na).The influence of the nature of the transition metal and the polyanionic group corresponding to the network-forming element (PO4, BO3, SiO4) on the electrical properties of the as-prepared glasses measured at different temperatures will be presented mainly for the lithiated compounds. For microstructural characterization, several techniques were coupled such as X-Ray Diffraction (XRD), SEM-EDX, DTA/TGA, Raman, IR and UV-Vis spectroscopies, Electrochemical Impedance Spectroscopy (EIS) and chronoamperometry.Finally, the electrochemical properties in terms of specific capacity, redox potentials, first cycle capacity loss, coulombic and energetic efficiencies of these materials were investigated in coin-cells vs lithium (or sodium) metal electrode. To understand the electrochemical processes involved in these materials at room temperature, SEM-EDX, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and Mossbauër Spectroscopy for iron based compounds were coupled and performed on the as-prepared glasses and on ex-situ (after cycling) materials at different electrochemical states of charge.All the generated data and measurements were analyzed by machine learning approach and tools and in some cases compared with a more classical Design of Experiments (DOE) approach. For instance, figure 1 illustrates the contour diagram of first discharge capacities in the mixture diagram of phosphate glasses based on 3 different transition metal oxides. The final goal is to establish more precisely the relationships between the different glassy oxides structures and their electrochemical performances. First results of this important work of elaboration, characterization and data analysis will be presented here. This new study will bring significant elements to optimize both the glass composition and its elaboration conditions to obtain high performance cathode materials without critical materials.[1] Nitta, N. et al., Li-ion battery materials: present and future. Materials Today 18, 252–264 (2015).[2] Tesfamhret, Y. et al., On the Manganese Dissolution Process from LiMn2O4 Cathode Materials. ChemElectroChem 8, 1516–1523 (2021).[3] K. Kercher et al., « Mixed polyanion glass cathodes: Glass-state conversion reactions », J. Electrochem. Soc., vol. 163, no 2, p. A131-A137, 2016, doi: 10.1149/2.0381602jes [4] Afyon, S. et al., New High Capacity Cathode Materials for Rechargeable Li-ion Batteries. Scientific Reports 4, 7113 (2015).[5] Kindle, M. et al., Alternatives to Cobalt. ACS Sustainable Chemistry Engineering 9, 629–638 (2021). Figure 1