Abstract:Solid polymer electrolytes are an exciting solution for safe and stable solid lithium electrode battery systems but are hindered by low ionic conductivity and low lithium transference. All-atom molecular dynamics simulation has become an invaluable tool to probe lithium diffusion mechanisms and accelerate the discovery of promising polymer chemistries. Because of their low computational cost and despite their approximate nature, only classical interatomic potentials can access the time and length scales for appropriate statistics of polymer kinetics. Machine learning (ML) potentials trained end-to-end on ab initio data have proven more accurate but cannot be scaled to the necessary time- and length- scales yet. Historical approaches to parameterize classical force fields have been incremental, reliant on a manual combination of top-down and bottom-up fitting, and are often paywalled and hard to reproduce. We introduce a computational learning workflow to predict classical interatomic potential parameters using quantum mechanical computations as training data that combines the automation and end-to-end fitting of ML with traditional class 1 and class 2 functional forms. The fitting strategy produced potentials whose simulations improved the accuracy of lithium coordination environments, diffusivities, and conductivities relative to experimental approaches when compared to both naive and hand-tuned parameters for liquid and solid organic electrolyte systems. We show that chemistry-informed regularization is necessary to constrain predicted parameters in order to reproduce experimental solvation and kinetic properties. Finally, we explore the limitations of non-polarizable, fixed point-charge schemes in describing electrolyte anions and compare the effects of two alternative schemes to fit point-charge distributions. The two strategies result in distinct lithium coordination mechanisms and highlight that closest parity to DFT forces and energies does not correlate to correct trends with lithium salt concentration in kinetic and solvation properties for fixed-point-charge classical interatomic potentials.

Automatic Machine-Learning Potential Generation Scheme and Simulation Protocol for the LiGePS-type Superionic Conductors

Deep Potential generation scheme and simulation protocol for the Li10GeP2S12-type superionic conductors

Machine Learning-Accelerated First-Principles Study of Atomic Configuration and Ionic Diffusion in Li10GeP2S12 Solid Electrolyte

Atomistic modeling of bulk and grain boundary diffusion in solid electrolyte Li$_6$PS$_5$Cl using machine-learning interatomic potentials

Size Dependent Lithium-Ion Conductivity of Solid Electrolytes in Machine Learning Molcular Dynamics Simulations

Simulating Diffusion Properties of Solid‐State Electrolytes via a Neural Network Potential: Performance and Training Scheme

High-throughput investigation of stability and Li diffusion of doped solid electrolytes via neural network potential without configurational knowledge

Simulating diffusion properties of solid-state electrolytes via a neural network potential: Performance and training scheme

Universal Machine Learning Interatomic Potentials: Surveying Solid Electrolytes

Electronic charge density as a fast approach for predicting Li-ion migration pathways in superionic conductors with first-principles level precision

Analysis of diffusion in solid state electrolytes through MD-simulations, improvement of the Li-ion conductivity in β-Li3PS4 as an example

Machine Learning-Aided Discovery of Superionic Solid-State Electrolyte for Li-Ion Batteries

Investigating Ionic Diffusivity in Amorphous Solid Electrolytes using Machine Learned Interatomic Potentials

Molecular dynamics simulations of the Li-ion diffusion in the amorphous solid electrolyte interphase

Machine-learning Assisted High-Throughput Discovery of Solid-State Electrolytes for Li-ion Batteries

A Pre-trained Deep Potential Model for Sulfide Solid Electrolytes with Broad Coverage and High Accuracy

Tackling Structural Complexity in Li2S-P2S5 Solid-State Electrolytes Using Machine Learning Potentials

On the role of long-range electrostatics in machine-learned interatomic potentials for complex battery materials

End-To-End Learning of Classical Interatomic Potentials for Benchmarking Anion Polarization Effects in Lithium Polymer Electrolytes

Mechanistic Insights into Temperature Effects for Ionic Conductivity in Li6PS5Cl