Predicting Oxidation and Spin States by High-Dimensional Neural Networks: Applications to Lithium Manganese Oxide Spinels

Abstract

Lithium ion batteries often contain transition metal oxides like Li${}_x$Mn${}_2$O${}_4$ ($0\le x\le 2$). Depending on the Li content different ratios of Mn${}^{\text{III}}$ to Mn${}^{\text{IV}}$ ions are present. In combination with electron hopping the Jahn-Teller distortions of the Mn${}^{\text{III}}$O${}_6$ octahedra can give rise to complex phenomena like structural transitions and conductance. While for small model systems oxidation and spin states can be determined using density functional theory (DFT), the investigation of dynamical phenomena by DFT is too demanding. Previously, we have shown that a high-dimensional neural network potential can extend molecular dynamics (MD) simulations of Li${}_x$Mn${}_2$O${}_4$ to nanosecond time scales, but these simulations did not provide information about the electronic structure. Here we extend the use of neural networks to the prediction of atomic oxidation and spin states. The resulting high-dimensional neural network is able to predict the spins of the Mn ions with an error of only 0.03 $\hslash$. We find that the Mn e${}_{\text{g}}$ electrons are correctly conserved and that the number of Jahn-Teller distorted Mn${}^{\text{III}}$O${}_6$ octahedra is predicted precisely for different Li loadings. A charge ordering transition is observed between 280 and 300 K, which matches resistivity measurements. Moreover, the activation energy of the electron hopping conduction above the phase transition is predicted to be 0.18 eV deviating only 0.02 eV from experiment. This work demonstrates that machine learning is able to provide an accurate representation of both, the geometric and the electronic structure dynamics of Li${}_x$Mn${}_2$O${}_4$, on time and length scales that are not accessible by ab initio MD.

arXiv comment15 pages, 12 figures

References in corpus (13)

• Global Optimization by Basin-Hopping and the Lowest Energy Structures of Lennard-Jones Clusters Containing up to 110 Atoms
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• Machine learning for molecular simulation
• Machine Learning of Molecular Electronic Properties in Chemical Compound Space
• How van der Waals interactions determine the unique properties of water
• Resolving transition metal chemical space: feature selection for machine learning and structure-property relationships
• Interatomic potentials for ionic systems with density functional accuracy based on charge densities obtained by a neural network
• Machine Learning for Quantum Mechanical Properties of Atoms in Molecules
• Predicting Electronic Structure Properties of Transition Metal Complexes with Neural Networks
• Electronic structure of Pr${}_{1-x}$Ca${}_x$MnO${}_3$
• Closing the gap between theory and experiment for lithium manganese oxide spinels using a high-dimensional neural network potential
• A Hybrid Density Functional Theory Benchmark Study on Lithium Manganese Oxides
• A criticial view on e${}_g$ occupancy as a descriptor for oxygen evolution catalytic activity in LiMn${}_2$O${}_4$ nanoparticles

Cited by in corpus (11)

• A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry
• High-Dimensional Neural Network Potentials for Magnetic Systems Using Spin-Dependent Atom-Centered Symmetry Functions
• Tutorial: How to Train a Neural Network Potential
• Insights into lithium manganese oxide-water interfaces using machine learning potentials
• Lifelong Machine Learning Potentials
• Machine Learning for Exploring Small Polaron Configurational Space
• Representations and Strategies for Transferable Machine Learning Models in Chemical Discovery
• Ab initio machine learning of phase space averages
• CoRe Optimizer: An All-in-One Solution for Machine Learning
• Atomistic Simulations of Oxide-Water Interfaces using Machine Learning Potentials
• Teaching oxidation states to neural networks