Multiscale Computational Screening of Metal–Organic Frameworks for Kr/Xe Adsorption Separation: A Structure–Property Relationship-Based Screening Strategy

The separation of radioactive noble gases, such as Xe and Kr, has attracted special attention in the context of used nuclear fuel (UNF). In this study, 180 metal-organic frameworks (MOFs) formally used for selective adsorptions of ethane and ethylene, with a similar kinetic diameter to Kr and Xe, were initially screened for the Kr/Xe separation using the grand canonical Monte Carlo (GCMC) method. Then, the structure-adsorption property relationships were generalized, that is, the MOFs of higher Kr/Xe selectivity are with the porosity at 0.2-0.4 and the ratio of the largest cavity diameter/pore limiting diameter at 1.0-2.4. Based on the relationships, six reported MOFs with large Kr uptakes and Kr/Xe selectivities were experimentally screened out and validated by GCMC simulations within the CoRE-MOF database, which are higher than most reported MOFs under conditions pertinent to nuclear fuel reprocessing of an 80/20 v/v mixture of Kr/Xe at normal temperature and pressure. Further simulations reveal that higher Kr uptakes and Kr/Xe selectivities of six MOFs result from the confinement effect of the pores. Molecular dynamic simulations showed that the six MOFs are ideal membrane separation materials of Kr from Xe, which are driven by adsorption and diffusion. Analyses of electronic structure-based density functional theory calculations showed that the main interaction between Kr and the six MOFs is van der Waals force dominated by dispersion and induction interactions. Therefore, the generalized structure-adsorption property relationships may assist the screening of MOFs for the separation and production of Kr/Xe from UNF industrially.