High-throughput computational screening of 137953 metal-organic frameworks for membrane separation of a CO2/N2/CH4 mixture

As an environmentally benign energy source, natural gas needs to be upgraded by separating impurities (e.g. CO2 and N-2) from CH4. In the current upgrading technology, CO2 and N-2 are separated via multiple energyintensive steps. Herein, we report a computational study to high-throughput screen 137953 metal-organic frameworks (MOFs) for a single-step membrane separation of a CO2/N-2/CH4 mixture. The screening strategy consists of four stages. First, the pore limiting diameters (PLDs) are calculated and 17257 MOFs with a PLD of 3-4 angstrom are selected. Second, the Henry's constants, diffusivities and permeabilities of CO2, N-2, and CH4 in the 17257 MOFs are estimated at 298 K and infinite dilution by Monte Carlo and molecular dynamics simulations. Their quantitative relationships with the PLD are established. Third, based on permselectivity versus permeability plots, 24 MOFs are prescreened for both CO2/CH4 and N-2/ CH4 separation. Finally, the adsorption, diffusion, and permeation of a three-component CO2/N-2/CH4 mixture are simulated in the 24 prescreened MOFs at 298 K and 10 bar, and the 5 best MOFs are identified for the membrane separation of CO2 and N-2 from CH4. This computational study reveals that the PLD and a new structural parameter (the percentage of pore size distribution) are key factors governing diffusion and permeation, provides quantitative structure-property relationships from the bottom up and would facilitate the development of new membranes for the upgrading of natural gas.