Multiple ensemble graph neural networks for the high-throughput screening of ionic liquids for carbon capture: Modeling study and experimental validation

Key Points

• High-throughput screening identifies ionic liquids with promising carbon capture capabilities.
• The top-performing ionic liquid demonstrated a carbon capture efficiency of 78% in experimental validation.
• Modeling study employed multi-ensemble graph neural networks to predict ionic liquid performance.
• Findings highlight the potential of machine learning methods to optimize carbon capture materials.