Critical Factors in Computational Characterization of Hydrogen Storage in Metal-Organic Frameworks

Inconsistencies in high-pressure H2 adsorption data and a lack of comparative experimenttheory studies have made the evaluation of both new and existing metalorganic frameworks (MOFs) challenging in the context of hydrogen storage applications. In this work, we performed grand canonical Monte Carlo (GCMC) simulations in nearly 500 experimentally refined MOF structures to examine the variance in simulation results because of the equation of state, H-2 potential, and the effect of density functional theory structural optimization. We find that hydrogen capacity at 77 K and 100 bar, as well as hydrogen 100-to-5 bar deliverable capacity, is correlated more strongly with the MOF pore volume than with the MOF surface area (the latter correlation is known as the Chahines rule). The tested methodologies provide consistent rankings of materials. In addition, four prototypical MOFs (MOF-74, CuBTC, ZIF-8, and MOF-5) with a range of surface areas, pore structures, and surface chemistries, representative of promising adsorbents for hydrogen storage, are evaluated in detail with both GCMC simulations and experimental measurements. Simulations with a three-site classical potential for H-2 agree best with our experimental data except in the case of MOF-5, in which H-2 adsorption is best replicated with a five-site potential. However, for the purpose of ranking materials, these two choices for H-2 potential make little difference. More significantly, 100 bar loading estimates based on more accurate equations of state for the vaporliquid equilibrium yield the best comparisons with the experiment.