UC Berkeley

Spinel-based materials have historically been one of the most promising frameworks for Li-ion cathodes since the discovery of LiMn2O4 in 1983. The material was first proposed due to the diamond network of face-shared unoccupied tetrahedral and octahedral sites in Mn2O4, which offers a three-dimensional pathway for Li-ion transport. Since then, much work has been spent on LiMn2O4 and its Ni-substituted derivative due to their use of Mn, which is more abundant and less expensive than the Co used in the layered-based Li-ion cathodes that dominate the portable electronic and electric vehicle industries. Furthermore, the three-dimensional Li-ion conduction pathway in spinel has made the spinel framework appealing for more applications than just Li-ion cathodes. Various spinel oxides and sulfides have been proposed as potential Mg-ion cathodes, and a low voltage Ti-based spinel has been proposed as a "zero-strain" Li-ion anode. The addition of spinel-like partial order has also been proposed to aid the voltage profile and Li transport in the disordered rock salt materials, which are a more recently-discovered class of high energy density Li-ion cathodes.

In this dissertation, I utilize first-principles techniques based on density functional theory calculations and cluster expansion methods to investigate intercalation in a number of promising spinel-based electrodes. In Chapter 1, I introduce in detail the spinel structure and the possible ways working ions can intercalate in the structure, as well as the techniques that will be used in the following studies.

In Chapter 2, I investigate the barriers to Mg intercalation into the spinel-MgxCr2O4 system. Using first-principles calculations in combination with a cluster expansion model and the nudged elastic band theory, I calculate the voltage curve for Mg insertion at room temperature and the activation barriers for Mg diffusion at varying Mg concentrations in the Cr2O4 structure. Our results identify a potential limitation to Mg intercalation in the form of stable Mg-vacancy orderings in the Cr2O4 lattice, which exhibit high migration barriers for Mg diffusion in addition to a steep voltage change. Additionally, we propose cation substitution as a potential mechanism that can be used to suppress the formation of stable Mg-vacancy orderings, which can eventually enable the practical usage of Cr2O4 as a Mg-ion cathode.

In Chapter 3, I delve into understanding the kinetics of Li transport in spinel Li4Ti5O12, which is a Li-ion anode exhibiting extraordinary rate capability. The rate capability of Li4Ti5O12 is especially surprising given its two-phase reaction and slow diffusion within the endmembers, in contrast to the typical fast-charging battery electrodes that are capable of accommodating lithium continuously via solid-solution transformation. Through real-time tracking of Li+ migration using operando electron energy-loss spectroscopy (EELS) combined with first-principles investigation of the Li configurations, their simulated EELS spectra, and Li migration barriers, it is revealed that the kinetic pathway that enables facile ionic transport in lithium titanate consists of distorted Li polyhedra in metastable intermediate states. This study highlights that the rate capability of fast-charging electrodes may not be controlled solely by the intrinsic ionic diffusivity of macroscopic phases, but also by transport via kinetically accessible low-energy landscapes. These findings may open new avenues for designing fast-charging electrodes.

In Chapter 4, I investigate the thermodynamic effects of disordering the spinel framework of LiMn2O4 towards understanding the partially disordered Li-excess spinels, a promising class of high energy density Li-ion cathodes with good discharge rate capability. To elucidate where in the partially disordered space the optimal materials lie, we investigate the configuration space of spinel with Mn disordered over the 16c and 16d sites. Using a cluster expansion in conjunction with Monte Carlo simulations, we find that disorder shortens the deleterious ~3V plateau between LiMn2O4 and Li2Mn2O4 by raising the energy landscape and stabilizing the motifs found in solid-solution configurations, and the plateau is completely removed with 25% Mn 16c occupancy. We expect partially disordered spinels with this level of disorder to exhibit both solid-solution behavior as well as good rate due to the remaining spinel character. We also identify Ti-doping and Li-excess as potential ways to facilitate 16c/16d disordering.