Rasmus Christensen

Disordered materials such as glasses are promising candidates as solid-state electrolytes for batteries, but they tend to exhibit low ionic conductivity and current methods for investigating and predicting the ionic conductivity of electrolyte materials are inefficient. Here, we propose a framework that summarizes the topological properties of the electrolyte material’s atomic structure with a weighted graph based on periodic boundary conditions. We then compute a variation of the max-flow algorithm, a commonly used model in continuum models of transportation and liquid flow, on this summary and show that it is compatible with the simulated periodic atomic structures. Using molecular dynamics simulations of various glassy solid-state electrolyte families including both mobile lithium and sodium ions, we validate the approach by showing a positive correlation between the area-normalized maximum flow predicted from the static glass structures and the ionic conductivity as obtained from the simulations, especially in the case of high-conductivity sulfide glass electrolytes. We envision that the developed methodology, which only relies on static atomic structure, will accelerate the discovery of highly conductive sulfide glassy electrolytes for high-performance all-solid-state batteries.

Diffusion in NaFePO ₄ can be enhanced through amorphization. Based on computations using DFT and machine learning potentials, we ascribe this phenomenon to the formation of less constrained Na-ion environments upon disordering.

Glass materials are promising candidates as solid electrolytes for batteries. However, the atomistic origins of variations in ionic conductivity with composition remain poorly understood. A key challenge in computationally designing high-performance glass electrolytes is the absence of an efficient general interatomic potential, particularly for mixed-anion systems, such as combinations of oxides, sulphides, and halides. To address these limitations, a machine learning interatomic potential for glass electrolytes is introduced, based on the atomic cluster expansion descriptor, covering common network formers, modifiers, and anions. Excellent agreement with experimental structure and property data across a wide range of glass compositions is observed. While the potential is broadly applicable to, e.g., lithium-ion conductivity, focus is here on the influence of glass composition on sodium-ion conductivity given the emerging importance of these conductors. By screening the diffusivity of 1,303 glass compositions, the presence of anion species, such as sulphides and halides, is found to have a large positive influence on diffusivity, enabling prediction of highly conductive glasses with mixed anions. Diffusivity and ionic conductivity are highly correlated with a set of descriptors, such as the average atomic volume. This work thus helps to accelerate predictive modelling and broader understanding of conductivity in glass electrolytes.

Most modern battery technologies depend on solid-state crystalline cathode materials. However, some of these materials are constrained by the low ionic conductivity of their most stable phases. An example of this is maricite (NaFePO4). Interestingly, experiments have shown that maricite can improve its rate capability through disordering (amorphization). However, experimental characterization of amorphous cathode materials remains a major challenge, hindering a clear understanding of the structural origin of the disorder-induced improvement in sodium-ion mobility. Here, we employ molecular dynamics simulations by first training a machine learning potential for NaFePO4 based on the atomic cluster expansion approach and a batch active learning potential parameterization scheme. This potential is then applied to explore the structural and dynamical properties of NaFePO4 glasses as cathode materials. Specifically, we investigate the effect of glass structure on sodium-ion diffusion, revealing the relative influences of short-range and medium-range order features. We find significant heterogeneity in sodium-ion diffusivity in the glass, with fast-conducting ions residing in less constrained atomic environments with fewer P and Fe neighbors. These more mobile ions are also surrounded by larger ring-type structures. Overall, the results and developed approach present promising avenues for developing high-performance glassy cathodes for next-generation batteries.

Understanding the dynamics of atoms in glasses is crucial for unraveling the origin of relaxation and the glass transition as well as predicting transport properties. However, identifying the structural features controlling atom dynamics in glasses remains challenging. Recently, machine learning models based on graph neural networks (GNNs) have successfully been used to predict future dynamics, but these prior studies focused primarily on model systems such as Kob–Andersen-type Lennard–Jones mixtures. This study investigates the use of local descriptors, GNN models, and molecular dynamics simulations to clarify the atomics dynamics in a realistic glass system (sodium silicate) across varying time scales. By harnessing the capabilities of different structural representations, we develop effective models for predicting the dynamics of sodium ions within the glassy silicate network, based solely on the initial atom positions. We further demonstrate the viability of our approach through comparison to previously proposed methods. Our findings pave the way for designing new glass formulations with tailored dynamical properties (e.g., as glassy electrolytes for batteries). Impact statement: Glass science has long grappled with understanding the fundamental nature and origin of glassy dynamics. The governing principles of atomic dynamics in glasses remain elusive as it is not obvious what to look for in the glass structure. While previous studies have focused on simplified model systems, we demonstrate for the first time that machine learning models can be used to accurately predict multi-time scale atomic dynamics in a complex oxide glass (sodium silicate) from the static atomic structure. By comparing different machine learning architectures, we establish that graph neural networks outperform conventional structural descriptors for dynamics prediction, with graph representations being able to effectively capture the complex multibody correlations that govern dynamics. Our findings show that the future dynamics in oxide glasses on time scales up to nanoseconds are at least partially encoded in the initial glassy configuration itself, showing that glassy dynamics is not a completely stochastic process. The capability to predict dynamics from structure has major implications as it could provide new tools for rational design of glassy materials with tailored dynamical properties and functionalities, possibly accelerating development of advanced glasses for applications in areas such as solid-state batteries and nuclear waste immobilization. Graphical

Methods to improve the fracture toughness of oxide glasses are needed since low fracture toughness is a major bottleneck for their applications. To overcome this, it is critically important to investigate the effect of both short- and medium-range structural features on fracture toughness. Recent work reported a record-high fracture toughness for a bulk lithium aluminoborate glass subjected to hot compression. Here, we further explore the structural origin of this high fracture toughness by subjecting different alkali aluminoborate glasses to hot compression. Through a combination of x-ray total scattering experiments and atomistic simulations, we find that hot compression causes significant changes to both the short- and medium-range order structure of the glasses, e.g., increased coordination numbers (CNs) of network forming species and decreased average size of ring-type structures. To this end, we reveal positive correlations between the pressure-induced increase in fracture toughness and (i) the increase in average CN of network forming species and (ii) the area of the first sharp diffraction peak in the structure factor. Our study thus improves the understanding of which structural features benefit intrinsic toughening of oxide glasses.

Classical molecular dynamics simulations of glassy materials rely on the availability of accurate yet computationally efficient interatomic force fields. The parameterization of new potentials remains challenging due to the non-convex nature of the accompanying optimization problem, which renders the traditional optimization methods inefficient or subject to bias. In this study, we present a new parameterization method based on particle swarm optimization (PSO), which is a stochastic population-based optimization method. Using glassy silica as a case study, we introduce two interatomic potentials using PSO, which are parameterized so as to match structural features obtained from ab initio simulations and experimental neutron diffraction data. We find that the PSO algorithm is highly efficient at searching for and identifying viable potential parameters that reproduce the structural features used as the target in the parameterization. The presented approach is very general and can be easily applied to other interatomic potential parameterization schemes.