Ionic Transport Predicting: A Machine Learning Dive into K 3 HCh (Ch: S, Se, and Te) for Next-Generation Solid State Batteries

View at publisher Summarise with VIRAT
Authors: Bouchra Kaima Mechraoui, Rabie Arar, Mohamed Halit, Zoulikha Hebboul
Cited by: 0
References: 45

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

In this study, we investigate the ionic transport behavior of potassium-based antiperovskites (K 3 HCh, Ch: S, Se, and Te) through a comprehensive machine learning framework. The workflow included 809 Inorganic electrolyte compounds, feature generation from elemental properties, selection of the top 50 features, and training and testing of several predictive models, and the XGBoost model achieves a mean absolute error (MAE) of approximately 0.45. To validate the model, we applied it to lithium-based antiperovskites (Li 3 HCh, Ch = S, Se, Te) at both high and low temperatures. The predicted ionic conductivities via the XGBoost model showed minimum errors of 7.34%, 2.65%, and 1.58% for Li 3 HS, Li 3 HSe, and Li 3 HTe, respectively, compared to experimental values. For potassium analogues, the predicted ionic conductivities at 100 °C were 7.64 × 10 –8, 3.5 × 10 –8, and 1.9 × 10 –7 S·cm –1 for K 3 HS, K 3 HSe, and K 3 HTe, respectively. In addition, Nudged elastic band (NEB) calculations confirm that K 3 HTe possesses a low activation energy of 0.38 eV, highlighting its potential for efficient ion transport.

Fields of Study & Subjects

Machine Learning in Materials Science Advanced Battery Materials and Technologies Thermal Expansion and Ionic Conductivity

Publication details

Journal / source
ACS Applied Energy Materials
Volume
9
Issue
7
Pages
4262–4275
Publication date
31 March 2026
Language
EN