Integration of density functional theory and machine learning for materials discovery in energy applications

Abstract

The global transition toward clean and renewable energy systems demands the rapid identification of novel functional materials for hydrogen production, photocatalysis, CO₂ reduction, and electrochemical energy storage. Traditional experimental trial and error approaches, combined with computationally expensive standalone density functional theory (DFT) calculations, have long constrained the pace of materials discovery. The convergence of machine learning (ML) and DFT has emerged as a transformative paradigm that accelerates the identification, screening, and validation of energy-relevant materials. This review critically surveys the integration of DFT and ML across the energy-materials landscape. Examine how DFT generated datasets, encompassing bandgap values, formation energies, adsorption energies, and electronic structure descriptors, serve as the foundation for training predictive ML models, and delineate the methodological pipeline that connects them. The roles of curated materials databases, including the Materials Project, the Open Quantum Materials Database (OQMD), and AFLOW, in enabling data-driven discovery are discussed alongside the workflow-automation frameworks that operationalise them. Applications spanning the hydrogen evolution reaction (HER), photocatalytic water splitting, CO₂ reduction, and lithium-ion battery materials are reviewed. Current challenges, including data quality and coverage, DFT accuracy limitations, model interpretability, out of distribution generalisation, and the reproducibility of several landmark large-scale demonstrations, are addressed, and future directions toward autonomous materials discovery platforms integrating artificial intelligence, robotics, and quantum computing are outlined together with their present-day limitations and motivating open problems. The DFT - ML framework has matured from a proof of concept into an important engine of energy materials discovery. DFT - ML integration enables accelerated discovery and screening of functional energy materials. Graph-based learning provides efficient prediction of key DFT-derived materials properties. Active learning couples ML screening with targeted DFT calculations for efficient materials exploration. High-throughput workflows and open databases establish scalable, data-driven materials discovery pipelines. Critical assessment identifies data, accuracy, interpretability, reproducibility, and generalisation challenges.