Advanced prediction of perovskite stability for solar energy using machine learning

• Published in: Solar energy 2024-08, Vol.278, p.112782, Article 112782
• Main Authors: Sudha Priyanga, G., Sampath, Santosh, Shravan, P.V., Sujith, R.N., Mohamed Javeed, A., Latha, G.
• Format: Article
• Language: English
• Subjects: Energy above Hull; Formation Energy; Machine learning; Perovskites; Thermodynamic stability
• ISSN: 0038-092X; 1471-1257
• DOI: 10.1016/j.solener.2024.112782

•A model has been built with only the elemental features of the composition of perovskite materials.•Using the ExtraTrees regressor algorithm we achieved a higher accuracy of 93.6%, 94.75%, and 98.41% in predicting Ef, Ehull, and TF, respectively.•We discovered 45 novel compositions of perovskite oxynitrides (ABO2N) and two novel compositions of perovskite oxides (ABO3) which are energetically, thermodynamically, and structurally stable.•Compounds with higher average ionic character and low electronegativity of A cation showcase the utmost stability due to their lowest formation energy.•We discovered 45 novel oxynitrides, in 38 compositions, the electronegativity of A exceeds that of B which form distorted cubic perovskite crystal structure, while in other 8 cases, B has greater electronegativity than A which tends to form ideal cubic perovskite crystal structure. In this work, we delve into the realm of perovskite materials with a comprehensive analysis on its structural and thermodynamic stability. Employing a machine learning approach, our study focuses on three important features for stability prediction such as formation energy (Ef), energy above hull (Ehull), and tolerance factor (TF). These features act as key indicators, allowing us to understand the intricate balance of energy and thermodynamic stability in perovskite structures for solar energy applications. We achieve this by training machine learning models on datasets generated computationally using DFT. Understanding the structural prediction of perovskite materials (ABX3, ABO3, ABO2X and ABOX2), whether thermodynamically stable or unstable, is critical for assessing their suitability for photovoltaic or photocatalytic applications. This study examines 14,199 mixed perovskite halides, oxides, and oxynitrides in order to determine the relationship between the aforementioned parameters and perovskite material composition. When compared to other machine learning models, using the ExtraTrees regression algorithm results in a higher accuracy of approximately 93.6 %, 94.75 %, and 98.41 % in predicting Ef, Ehull, and TF, respectively. The proposed method not only predicts Ef, Ehull, and TF, but it also aids in the discovery of new materials. We are particularly interested in ABO3 and ABO2N compositions from this perovskite family. We have come up with 306 stable perovskite oxides and 311 stable oxynitrides using our prediction. Among these, we discovered 45 novel compositions of perovskite oxynitrid