Open Copper Site MOFs (OCS-MOFs) Database

This dataset represents a curated collection of 27,592 metal-organic frameworks (MOFs) with open copper sites (OCS) specifically selected for their potential in CO₂/CH₄ separation, which is essential for biogas upgrading. The OCS-MOFs were sourced and refined from the ARC-MOF database, following a comprehensive screening process to ensure novelty, validity, and applicability. The structures underwent rigorous quality control, including elimination of duplicates using pymatgen and MOFid, validation of physical and chemical integrity using MOFChecker, and assessment of pore accessibility via Zeo++.

The dataset serves as the foundation for the multi-scale molecular simulations and machine learning models presented in our accompanying research. These models employ OCS-specialized force fields and diversity-transferable machine learning approaches to unveil structure-function relationships in MOFs, providing valuable insights into pore geometry and local chemical environments that optimize CO₂ adsorption capacity and selectivity.

For the accompanying machine learning pipline code, and tabulated data, please visit the GitHub repository here. https://github.com/xiaoyu961031/OCS-MOFs