A database of experimentally measured lithium solid electrolyte conductivities evaluated with machine learning

Abstract

The application of machine learning models to predict material properties is determined by the availability of high-quality data. We present an expert-curated dataset of lithium ion conductors and associated lithium ion conductivities measured by a.c. impedance spectroscopy. This dataset has 820 entries collected from 214 sources; entries contain a chemical composition, an expert-assigned structural label, and ionic conductivity at a specific temperature (from 5 to 873 °C). There are 403 unique chemical compositions with an associated ionic conductivity near room temperature (15–35 °C). The materials contained in this dataset are placed in the context of compounds reported in the Inorganic Crystal Structure Database with unsupervised machine learning and the Element Movers Distance. This dataset is used to train a CrabNet-based classifier to estimate whether a chemical composition has high or low ionic conductivity. This classifier is a practical tool to aid experimentalists in prioritizing candidates for further investigation as lithium ion conductors.

Oxide Ion-Conducting Materials Containing Tetrahedral Moieties: Structures and Conduction Mechanisms

“…Even though ML has been successfully applied for developing novel lithium-ion conductors − (e.g., Li 3.3 SnS 3.3 Cl 0.7 ) and oxide-ion conductors within the Ca-(Nb,Ta)-Bi-O system, , it still remains barely employed, most probably due to the limited number of suitable datasets available so far; therefore, few effective descriptors have been identified. In this sense, this Review proposes a potential dataset containing solid-state electrolytes based on tetrahedral units and different migration mechanisms that may be then screened by different ML methods with more descriptors related to site disorder and chemical bonding . Thus, it is expected that the applicability of AI and ML could increase as the computational performance and accessibility improve …”

“…In this sense, this Review proposes a potential dataset containing solid-state electrolytes based on tetrahedral units and different migration mechanisms that may be then screened by different ML methods with more descriptors related to site disorder and chemical bonding. 283 Thus, it is expected that the applicability of AI and ML could increase as the computational performance and accessibility improve. 282…”

Machine Learning‐Assisted Property Prediction of Solid‐State Electrolyte

“…[116][117][118] Hargreaves et al introduced a meticulously curated dataset centered on lithium-ion conductors, along with their respective conductivities determined through impedance spectroscopy. [119] The compilation includes 820 entries sourced from 214 different references. Each entry includes details on the chemical makeup, a label given by experts based on structural characteristics, and the ionic conductivity at a particular temperature.…”

Evaluation of solid electrolytes: Development of conventional and interdisciplinary approaches

“…Machine learning (ML) methods have the potential to significantly enhance the accuracy of materials investigation because of their capacity for identifying complex patterns concealed underneath high-dimensional data. [108,[138][139][140][141][142][143] The ML on the basis of the materials database has been predictable to offer a complex nonlinear relationship between the property/function and structure/composition of the proposed materials such as the ionic conductor. [138,[144][145][146][147] A model was proposed for forecasting the ionic conductivity by ML based on currently available experimental data.…”