Uncertainty-Aware Active Learning Reveals Reliability Limits in Lead-Free Halide Perovskite Screening

Key Points

• This research aims to evaluate the reliability of lead-free perovskite materials for use in solar cells while addressing the inaccuracies in current screening methods.
• Utilized active learning frameworks to navigate the large chemical space of lead-free perovskites.
• Implemented high-throughput density functional theory calculations.
• Analyzed the limitations of current screening methods in identifying stable materials.
• Identified critical reliability limits in lead-free perovskite screening processes.
• Demonstrated that systematic inaccuracies could significantly alter the discovery of stable materials.
• Provided insights into improving screening methodologies for future perovskite research.