Computational Discovery of Solid Electrolytes for Batteries: Interfacial Phenomena and Ion Mobility

Abstract

Solid-state batteries (SSBs) using a solid electrolyte (SE) and Li-metal anode are promising technologies that can increase energy density and minimize safety concerns for applications such as electric vehicles. Although the recent discovery of SEs with high ionic conductivity has advanced the prospects for realizing SSBs, additional study of these materials has unearthed several shortcomings (e.g., interfacial degradation). Thus, the discovery of alternative SE remains an important pursuit. This search has been slowed, however, by incomplete understanding of the elementary features that give rise to high ionic mobility and promote interfacial stability. In response, this dissertation focuses on several topics that are relevant for the advancement of SSBs: (1) stability and wettability at interfaces between a SE and metal anode, (2) fundamental understanding of ionic transport mechanisms in solids, and (3) the discovery of new SEs. These topics are investigated using first-principles calculations. Anti-perovskites (AP) are adopted as model SEs because they have shown promise for achieving high ionic conductivities while possessing simple structures that enable a comprehensive characterization of their properties. In addition, machine learning (ML) is employed to analyze trends in the computed data. Investigation of the Li3OCl/Li interface shows that an oxygen-terminated interface is the most stable. This interface exhibits strong interfacial bonds, suggesting good wettability by Li, low interfacial resistance, and potential for Li dendrite prevention. However, this strong interaction also locally shifts the electronic band edge positions, narrowing the bandgap by 30%. Nevertheless, the conduction band minimum remains more negative than the Li/Li+ potential, implying stability against charge injection from the anode. These calculations indicate a tradeoff between strong interfacial bonding/wettability and electrochemical stability. Next, the connections between ion mobility, thermodynamic stability, and symmetry-lowering lattice distortions are characterized across 36 model APs. Compounds with larger lattice distortions exhibit smaller percolating migration barriers because these distortions speed up migration along a subset of pathways. As larger distortions also correlate with reduced stability, realizing high ionic mobility requires balancing a mobility/stability tradeoff. Li3SeF, Na3SeF, Na3SBr, Na3SF, K3SeF, and K3SBr are identified as new compositions that balance this tradeoff. Differences in ion mobilities across Li, Na, and K based APs is rationalized in terms of differences in ion packing, vibrational frequency, polarizability, and ionic charge. Next, using data generated for alkali metal-based APs, ML is used to identify elementary features that correlate with ionic mobility. Lattice structure was found to have a greater influence in ion transport than do features based on chemical or electronic properties. For vacancy migration, the migration distance and bottleneck-size are the most important features: migration barriers decrease with shorter hops and with wider migration channels. Therefore, tuning the structure of a SE is the most effective scheme to improve ion mobility in these compounds. Finally, potential multivalent MV-ion SEs based on the AP structure are examined. SEs with compositions Mg3NAs, Ca3NAs, and Ca3PSb are identified as the most promising. These compounds are predicted to be thermodynamically stable, electronically insulating, stable in contact with metal anodes, and to have relatively low percolating barriers for ion migration. Due to their high formation energies, ionic defects should be introduced artificially. In total, this study enhances understanding of interfacial phenomena and ion transport in solid electrolytes, while also suggesting new materials. The ultimate goal is to accelerate the introduction of SSBs with improved safety and energy density.