Calix[4]arene Design for Enhanced Carbon Capture via Topological Learning

Solid direct air capture (S-DAC) technologies have the potential to remove pre-existing CO₂ from the atmosphere by taking advantage of hydrophobic materials that will retain CO₂ through physisorption and separating it from the abundant water molecules. Calix[n]arenes are highly tunable and versatile complexes with inherent hydrophobic cavities, making them promising candidates for S-DAC technologies. In this study, we investigated four classes of calix[4]arenes, which were functionalized symmetrically at various sites, using machine learning workflows based on molecular topology. The Topology-Guided Carbon Capture with Calix[4]arenes (TopoC³) model presented here leverages DFT data from an initial molecular subset together with persistent homology-based topological descriptors to enable targeted molecular discovery. Recognizing the strong dependence of CO₂ host-guest interactions on molecular structure, this study identified key structural and topological features that govern adsorption performance. These insights establish clear design criteria for calix[4]arenes as molecular candidates for CO₂ capture and provide a principled framework for further molecular engineering.