AI Frontiers: Accelerating Material Discovery for Supercapacitors

AI Frontiers: Accelerating Material Discovery for Supercapacitors This study presents a unified AI-powered computational pipeline leveraging MEGNet and attention-enhanced MAGNET models to screen over 22,000 crystal structures for high-performance supercapacitor electrode materials. DFT-predicted properties — formation energy, hull energy, and band gap — combined with Monte Carlo Dropout ensure reliable candidate selection. Interpretable ML applied across MXenes, metal oxides, and carbon electrodes yielded 45 of 50 shortlisted candidates as high-capacitance materials, with V₂CTₓ and Ti₃C₂Tₓ MXenes as standout prospects. The work establishes an open, reproducible discovery paradigm for next-generation energy storage research.