From Formability to Bandgap: Machine Learning Accelerates the Discovery and Application of Perovskite Materials

Area of Science:

• Materials Science
• Computational Chemistry
• Artificial Intelligence

Background:

• Perovskite materials exhibit unique properties suitable for solar cells, photodetectors, catalysts, and LEDs.
• Traditional research methods for perovskites are often slow and computationally demanding.
• Machine learning (ML) offers a powerful approach to accelerate perovskite discovery and optimization.

Purpose of the Study:

• To review the impact and methodologies of machine learning in perovskite material science.
• To analyze the trends, workflows, and algorithms of ML applications in perovskite research.
• To highlight ML's role in predicting perovskite properties and guiding practical applications.

Main Methods:

• Literature review and analysis of recent machine learning applications in perovskite research.
• Delineation of machine learning workflows and fundamental algorithms for perovskite materials.
• Analysis of perovskite structure-composition relationships to predict properties like formability and bandgap.

Main Results:

• Machine learning significantly accelerates the discovery and optimization of perovskite materials.
• ML accurately predicts key perovskite properties, including formability and bandgap.
• Four practical applications of ML in perovskite technology are presented, alongside challenges and future outlooks.

Conclusions:

• Machine learning is revolutionizing perovskite research by enabling faster discovery and property prediction.
• This review provides insights and guidance for leveraging ML in perovskite development.
• Further integration of ML promises to enhance the practical applications of perovskite materials.