Abstract:Next generation batteries based on lithium (Li) metal anodes have been plagued by the dendritic electrodeposition of Li metal on the anode during cycling, resulting in short circuit and capacity loss. Suppression of dendritic growth through the use of solid electrolytes has emerged as one of the most promising strategies for enabling the use of Li metal anodes. We perform a computational screening of over 12 000 inorganic solids based on their ability to suppress dendrite initiation in contact with Li metal anode. Properties for mechanically isotropic and anisotropic interfaces that can be used in stability criteria for determining the propensity of dendrite initiation are usually obtained from computationally expensive first-principles methods. In order to obtain a large data set for screening, we use machine-learning models to predict the mechanical properties of several new solid electrolytes. The machine-learning models are trained on purely structural features of the material, which do not require any first-principles calculations. We train a graph convolutional neural network on the shear and bulk moduli because of the availability of a large training data set with low noise due to low uncertainty in their first-principles-calculated values. We use gradient boosting regressor and kernel ridge regression to train the elastic constants, where the choice of the model depends on the size of the training data and the noise that it can handle. The material stiffness is found to increase with an increase in mass density and ratio of Li and sublattice bond ionicity, and decrease with increase in volume per atom and sublattice electronegativity. Cross-validation/test performance suggests our models generalize well. We predict over 20 mechanically anisotropic interfaces between Li metal and four solid electrolytes which can be used to suppress dendrite growth. Our screened candidates are generally soft and highly anisotropic, and present opportunities for simultaneously obtaining dendrite suppression and high ionic conductivity in solid electrolytes.

Highly stable silicon oxycarbide all-solid-state batteries enabled by machined learning accelerated screening of oxides and sulfides electrolytes

A Dynamically Stable Sulfide Electrolyte Architecture for High-Performance All-Solid-State Lithium Metal Batteries

Machine Learning Approach Accelerates Search for Solid State Electrolytes

A High‐Throughput Search for Functionally Stable Interfaces in Sulfide Solid‐State Lithium Ion Conductors

Interfacial compatibility design of high-performance electrolytes for SiCO-based all-solid-state lithium battery

Metal–Organic Frameworks Derived Electrolytes Build Multiple Wetting Interfaces for Integrated Solid‐State Lithium–Oxygen Battery

A stable solid-state lithium-oxygen battery enabled by heterogeneous silicon-based interface

Engineering Electrode–electrolyte Interface for Ultrastable Si-based Solid-State Batteries

Weakly Solvated Electrolyte Enables the Robust Solid Electrolyte Interface on Siox Anodes for Lithium-Ion Battery

Machine Learning Enabled Computational Screening of Inorganic Solid Electrolytes for Suppression of Dendrite Formation in Lithium Metal Anodes

Constructing a Robust Solid-Electrolyte Interphase Layer via Chemical Prelithiation for High-Performance SiO<i><sub>x</sub></i> Anode

Probing Solid–Solid Interfacial Reactions in All-Solid-State Sodium-Ion Batteries with First-Principles Calculations

Constructing High Performance Dry-Processing Oxide Composite Electrolyte Via Interfacial Interactions for Durable Solid-State Lithium Batteries

Organic-inorganic multi-scale enhanced interfacial engineering of sulfide solid electrolyte in Li-S battery

In-situ Constructing 3D Nanocarbon Conduction Conformal Network on Silicon Anodes for High-Performance Sulfide-Based All-Solid-state Batteries

Rational Design of an In-Build Quasi-Solid-state Electrolyte for High-Performance Lithium-Ion Batteries with the Silicon-Based Anode

Magnesio-mechanochemical Reduced SiOx for High-Performance Lithium Ion Batteries

Batch-Scale Synthesis and Interfacially Enhanced Stability of Silicon Suboxide-Based Anodes toward High-Performance Lithium-Ion Batteries

Computational approach inspired advancements of solid-state electrolytes for lithium secondary batteries: from first-principles to machine learning

Achieving Both High Ionic Conductivity and High Interfacial Stability with the Li2+xC1-xBxO3 Solid-State Electrolyte: Design from Theoretical Calculations

Electrolyte Design Enabling Stable Solid Electrolyte Interface for High-Performance Silicon/Carbon Anodes