Mechanistic study of mixed lithium halides solid state electrolytes

TL;DR

This work investigates how halide alloying and metal substitution affect structure and Li-ion conductivity in Li$_3$YCl$_{6x}$Br$_{6(1-x)}$ using a universal MLIP (PET-MAD) trained on a diverse atomic dataset. The authors quantify phase stability, halide disorder, and transport properties with MD/MC and Green-Kubo calculations, finding that halide distribution is effectively random and that conductivity is governed by a balance between lattice contraction (driven by Cl substitution) and shorter Y–X bonds that facilitate diffusion. They demonstrate that the 1:1 Br/Cl composition often yields the highest qualitative conductivity and that constant-volume vs constant-pressure analyses reveal compensating effects between volume and composition, providing design rules for halide SSEs. Indium substitution mirrors the yttrium results, indicating generalizability of the approach across metal substitutions. Although absolute conductivities are overestimated relative to experiments, the study delivers mechanistic insights and a practical MLIP-based framework for accelerating discovery of high-performance halide solid-state electrolytes.

Abstract

Lithium halides with the general formula Li$_x$M$_y$X$_6$, where M indicates transition metal ions and X halide anions are very actively studied as solid-state electrolytes, because of relatively low cost, high stability and Li conductivity. The structure and properties of these halide-based solid electrolytes (HSE) can be tuned by alloying, e.g. using different halides and/or transition metals simultaneously. The large chemical space is difficult to sample by experiments, making simulations based on broadly applicable machine-learning interatomic potentials (MLIPs) a promising approach to elucidate structure-property relations, and facilitate the design of better-performing compositions. Here we focus on the Li$_3$YCl$_{6x}$Br$_{6(1-x)}$ system, for which reliable experimental data exists, and use the recently-developed PET-MAD universal MLIP to investigate the structure of the alloy, the interplay of crystalline lattice, volume and chemical composition, and its effect on Li conductivity. We find that the distribution of Cl and Br atoms is only weakly correlated, and that the primary effect of alloying is to modulate the lattice parameter -- although it can also trigger transition between different lattice symmetries. By comparing constant-volume and constant-pressure simulations, we disentangle the effect of lattice parameter and chemical composition on the conductivity, finding that the two effects compensate each other, reducing the overall dependency of conductivity on alloy composition. We also study the effect of Y-In metal substitution finding a small increase in the conductivity for the C2/m phase at 25\% In content, and an overall higher conductivity for the P$\bar{3}$m1 phase.

Figures (9)

• Figure 1: Formation energy for the P$\bar{3}$m1 and C2/m structures as a function of the Cl content, computed with both the PET-MAD potential and its fine-tuned version. For both potentials, the baseline energies are those of the C2/m phase. For the mixed composition, the results shown are the average values of the energies of the 4 structures selected from the MC, and the shaded areas represent the semi-dispersion of the results.
• Figure 2: Distribution of octahedral compositions in the first coordination shell of Y atoms during the MC simulation for the P$\bar{3}$m1 (first column) and C2/m (second column) structures. Black stars indicate the expected values from a perfect binomial distribution, with $p$ corresponding to the ratio between the number of Cl and the total number of halogen atoms in the formula units: $0.5$ for the first row, $0.25$ for the second, and $0.75$ for the third. The heights of the bars and the distributions are normalized such that the total area sums to 1.
• Figure 3: Radial distribution functions (RDF) of a configuration of Li$_3$YBr$_3$Cl$_3$ in the P$\bar{3}$m1 phase (left) and C2/m (right) for both the PET-MAD (blue) and the finetuned potential (orange). The upper panels contain the RDF between the Y atoms and the halogens: Y-Br (continuous), Y-Cl (dashed). The lower panels contain the RDF between the different halogen atoms: Br-Br (continuous), Br-Cl (dotted), Cl-Cl (dashed). These are extracted from the final $50$ ps of $2$ ns long NPT simulations.
• Figure 4: Y-Y RDF for the P$\bar{3}$m1 (a), C2/m (b) phases for the different chemical compositions. Behavior of the volume per formula unit (c) and the position of the first peak in the Y-Y RDF (b) as a function of the chemical composition.
• Figure 5: Ionic conductivity ($\sigma$) of Li$3$YBr$_{6(1-x)}$Cl$_6x$, with $x\in[0,1]$, as a function of percentage of Cl content, computed via the PET-MAD universal potential (blue) and a PET model fine-tuned (pink line) on a specific dataset. The shaded area indicates the statistical uncertainty for the pure Li3YBr6 and Li3YCl6 phases; for the mixed composition, it indicates the semi-dispersion among the 4 from the different simulations selected during the MC procedure. Results are compared with experimental results reported in Refs. Liu2021Asano2018 (green) and Ref. Maas2023 (brown).