Machine Learning Enabled Computational Screening of Inorganic Solid Electrolytes for Dendrite Suppression with Li Metal Anode (1804.04651v1)

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 dataset for screening, we use machine learning models to predict the mechanical properties of several new solid electrolytes. We train a convolutional neural network on the shear and bulk moduli purely on structural features of the material. We use AdaBoost, Lasso and Bayesian ridge regression to train the elastic constants, where the choice of the model depended on the size of the training data and the noise that it can handle. Our models give us direct interpretability by revealing the dominant structural features affecting the elastic constants. The stiffness is found to increase with a decrease in volume per atom, increase in minimum anion-anion separation, and increase in sublattice (all but Li) packing fraction. Cross-validation/test performance suggests our models generalize well. We predict over 20 mechanically anisotropic interfaces between Li metal and 6 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.

Overview of Machine Learning Enabled Computational Screening of Inorganic Solid Electrolytes for Dendrite Suppression with Li Metal Anode

The paper by Ahmad et al. focuses on a computational approach to address the persistent challenge of dendrite formation in lithium (Li) metal anodes, which significantly impedes the performance and safety of Li-based batteries. The work utilizes ML models to predict mechanical properties of a large database of inorganic solid electrolytes. This prediction serves as the basis for evaluating their potential to suppress dendrite initiation when in contact with Li metal.

Key Results

The paper performs an extensive screening of over 12,000 inorganic solids, leveraging ML tools to estimate mechanical properties without resorting to computationally expensive first-principles calculations. A convolutional neural network (CNN) is employed to predict shear and bulk moduli, while additional methods such as AdaBoost and Lasso regression are used for training elastic constants. These models are trained using structural features of the materials and are validated through cross-validation metrics, suggesting robust generalization capabilities.

Anisotropic interfaces, characterized by distinctive crystallographic orientations, are emphasized due to the anisotropic properties of Li metal itself. This is critical as grain orientation strongly affects mechanical compatibility and thus dendrite suppression capability. The analysis identifies over twenty promising interfaces between Li and specific solid electrolytes, hypothesized to enhance stability against dendritic growth.

Implications and Future Directions

This paper provides a blueprint for accelerating the discovery of materials with potential for dendrite suppression in the context of solid-state batteries. The use of ML models highlights the ability to bypass the computational bottleneck associated with traditional simulation methods, enabling rapid screening and facilitating the discovery of novel material candidates.

The implications of this work are significant for the development of next-generation Li batteries with enhanced safety and performance metrics. The focus on mechanically soft yet anisotropic materials also opens up pathways for optimizing mechanical properties alongside ionic conductivity. Such optimization is crucial for practical applications where both stability and efficiency are required.

In terms of theoretical implications, the paper underscores the importance of anisotropic mechanical treatments over isotropic assumptions typically employed. The complex interplay between crystallographic orientation and mechanical properties is pivotal in the stabilization of Li-anode interfaces.

Future developments could expand the dataset of materials to include non-cubic symmetries, as well as glassy or amorphous structures, which are often overlooked despite their potential in electrolyte applications. Additionally, the integration of experimental validation with the ML predictions could establish a feedback loop for continuous refinement of models, further improving prediction accuracy.

Conclusion

Ahmad et al.'s paper illustrates an innovative approach to addressing the dendrite problem in Li metal anodes through machine learning-enabled screening of solid electrolytes. By identifying promising material candidates that prevent dendrite initiation, this research lays foundational work towards developing safer and more reliable Li-based batteries. The methods and insight provided offer a clear direction for subsequent research efforts in both computational material science and solid-state electrochemistry.