Combined GCMC, MD, and DFT Approach for Unlocking the Performances of COFs for Methane Purification

Abstract

Covalent organic frameworks (COFs) are promising materials for gas storage and separation; however, the potential of COFs for separation of CH₄ from industrially relevant gases such as H₂, N₂, and C₂H₆ is yet to be investigated. In this work, we followed a multiscale computational approach to unlock both the adsorption- and membrane-based CH₄/H₂, CH₄/N₂, and C₂H₆/CH₄ separation potentials of 572 COFs by combining grand canonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations and density functional theory (DFT) calculations. Adsorbent performance evaluation metrics of COFs, adsorption selectivity, working capacity, regenerability, and adsorbent performance score were calculated for separation of equimolar CH₄/H₂, CH₄/N₂, and C₂H₆/CH₄ mixtures at vacuum swing adsorption (VSA) and pressure swing adsorption (PSA) conditions to identify the best-performing COFs for each mixture. Results showed that COFs could achieve selectivities of 2-85, 1-7, and 2-23 for PSA-based CH₄/H₂, CH₄/N₂, and C₂H₆/CH₄ separations, respectively, outperforming conventional adsorbents such as zeolites and activated carbons for each mixture. Structure-performance relations revealed that COFs with pore sizes <10 Å are promising adsorbents for all mixtures. We identified the gas adsorption sites in the three top-performing COFs commonly identified for each mixture by DFT calculations and computed the binding strength of gases, which were found to be on the order of C₂H₆ > CH₄ > N₂ > H₂, supporting the GCMC results. Nucleus-independent chemical shift (NICS) indexes of aromaticity for adsorption sites were calculated, and the results revealed that the degree of linker aromaticity could be a measure for the selection or design of highly alkane-selective COF adsorbents over N₂ and H₂. Finally, COF membranes were shown to achieve high H₂ permeabilities, 4.57 × 10³ -1.25 × 10⁶ Barrer, and decent membrane selectivities, as high as 4.3, outperforming polymeric and MOF-based membranes for separation of H₂ from CH₄.

Figure 1

Selectivity and working capacity of 572 COFs for separation of (a, b) CH₄/H₂: 50/50 mixture, (c, d) CH₄/N₂: 50/50 mixture, and (e, f) C₂H₆/CH₄: 50/50 mixture at VSA (1–0.1 bar at 298 K) and PSA (10–1 bar at 298 K) conditions. Colors represent APSs of COFs. MOFs from our previous works^,, (black data points, 4350 MOFs for CH₄/H₂, 4738 MOFs for CH₄/N₂, and 281 MOFs for C₂H₆/CH₄ separations) were included in (b), (d), and (f) for comparison.

Figure 2

R% and APSs of 572 COFs for separation of (a, b) CH₄/H₂ mixture, (c, d) CH₄/N₂ mixture, and (e, f) C₂H₆/CH₄ mixture at VSA and PSA conditions. Colors represent the porosity of COFs, and stars represent the top 10 COFs identified for each condition. Red dashed line represents the R% = 85% line.

Figure 3

PLDs and porosities of 572 COFs as a function of their APSs calculated for (a) CH₄/H₂ and (b) CH₄/N₂ separation at PSA conditions, and (c, d) C₂H₆/CH₄ separation at VSA and PSA conditions. Big circles with thick edges represent the top 10 COFs identified for each mixture at each condition.

Figure 4

Three-dimensional representations and electrostatic potential maps of the linker fragments of the three common top COFs. White, hydrogen; gray, carbon; blue, nitrogen. Negative (positive) regions are depicted in red (blue) color.

Figure 5

(a) Calculated membrane, diffusion, and adsorption selectivities of 572 COFs at 1 bar. Hollow circles represent the top COF membranes identified for H₂/CH₄ separation. S_ads,CH4/H2 of COFs are provided in purple under the x-axis. (b) H₂ permeabilities and H₂/CH₄ membrane selectivities of COFs calculated at 1 bar along with Robeson’s upper bound. Colors represent the porosities of COFs. Membrane-based H₂/CH₄ separation performances of 4240 MOFs obtained from our previous work are given in gray diamonds for comparison.