Thermal regimes of Li‐ion conductivity in β‐eucryptite

Abstract

While it is well-established that ionic conduction in lithium aluminosilicates proceeds via hopping of Li ions, the nature of the various hoping-based mechanisms in different temperature regimes has not been fully elucidated. The difficulties associated with investigating the conduction have to do with the presence of grains and grain boundaries of different orientations in these usually polycrystalline materials. Herein, we use electrochemical impedance spectroscopy (EIS) to investigate the ion conduction mechanisms in b-eucryptite, which is a prototypical lithium aluminosilicate. In the absence of significant structural transitions in grain boundaries, we find that there are three conduction regimes for the one-dimensional ionic motion along the c axis channels in the grains, and determine the activation energies for each of these temperature regimes. Activation energies computed from molecular statics calculations of the potential energy landscape encountered by Li ions suggest that at temperatures below 440°C conduction proceeds via cooperative or correlated motion, in agreement with established literature. Between 440°C and 500°C, the activation barriers extracted from EIS measurements are large and consistent with those from atomistic calculations for uncorrelated Li ion hopping. Above 500°C the activation barriers decrease significantly, which indicates that after the transition to the Li-disordered phase of b-eucryptite, the Li ion motion largely regains the correlated character. K E Y W O R D Saluminosilicates, impedance spectroscopy, ionic conduction, lithium 1 | INTRODUCTION As a prototype of lithium aluminum silicates (LAS), b-eucryptite (LiAlSiO 4 ) has attracted both fundamental and technological interest for decades due to its phase transformations, 1-5 unusual thermo-mechanical properties, 6-10 and the one-dimensional nature of its ionic conduction. [11][12][13][14][15] The superior Li ion conduction makes b-eucryptite a promising candidate for electrolyte applications, including thermal batteries and high-temperature solid electrodes. [16][17][18] Furthermore, applications such as electrothermal devices as well as thermal shock resistant structures can benefit from the overall negative coefficient of thermal expansion (CTE) of b-eucryptite. 8,9,16,17,[19][20][21][22][23][24] The crystal structure of b-eucryptite (space group P6 4 22 or P6 2 22

Unsupervised landmark analysis for jump detection in molecular dynamics simulations

“…12 the Li-ion densities sampled during the dynamics. The unidimensional channels of ionic diffusion are compatible with published results [71,73]. The diffusion coefficient is hard to converge for the short dynamics we obtained for this system, and so quantifying the diffusion coefficient and its error cannot be done rigorously.…”

“…The structure of the β-eucryptite LiAlSiO 4 [66], referred to as LASO hereafter, is taken from COD [67] entry 9000368. It has been studied for its anisotropic expansion coefficient [68,69] and its ionic conductivity [70][71][72][73]. The structure can be described as an ordered β-quartz solid solution, with alternating aluminum and silicon planes.…”

Machine learning for multi-fidelity scale bridging and dynamical simulations of materials

“…Its generality, i.e., being applicable to a diverse range of materials (metals, ceramics, amorphous glasses, polymers, or biomolecules), along with its versatility, i.e, the ability to capture thermodynamic, mechanical, electrical and chemical behavior of materials, make it a pervasive and vital theoretical tool, as highlighted in Figure 1a. Increasingly, empirical findings are supported with analogous MD simulations to explain materials behavior at the electronic, atomistic or molecular level [12,13]. Beyond its scientific merit, the popularity of MD simulations is fueled by the advancements in the computational power, and the availability of efficient software packages (VASP [14], LAMMPS [15], ASE [16], NAMD [17], CHARMM [18], GROMACS [19], etc.)…”

Multi-reward Reinforcement Learning Based Bond-Order Potential to Study Strain-Assisted Phase Transitions in Phosphorene

“…More importantly, while there has been a significant amount of computational effort, most have primarily involved the use of first-principles calculations based on DFT. Dynamical studies that aim at understanding the structural and phase transformation upon application of external stimuli (such as mechanical, thermal, , etc.) over time (picoseconds to microseconds) and length scales (≫tens of nanometers) are of significant importance.…”