Metal-organic frameworks (MOFs) have emerged as an important class of porous materials that have the potential to be designed and functionalized for use in applications such as separations and catalysis. They have recently garnered much attention as possible materials for storing hydrogen or natural gas. MOFs are synthesized by a self-assembly process in which metal or metal-oxide vertices are connected by rigid or semi-rigid organic molecules. Because MOFs have predictable synthesis techniques and uniform framework structures, molecular modeling can be employed in a design capacity for screening hypothetical structures before they are synthesized. Molecular modeling can also be used to provide important information regarding adsorption properties that can aid in the characterization of novel structures. MOFs have been shown to have some of the highest surface areas and lowest crystal densities reported for any material to date. Surface areas are typically reported as BET surface areas, obtained from nitrogen adsorption isotherms at 77 K. However, the BET method relies on several assumptions that may break down for microporous materials with very large surface areas. For instance, the BET analysis is questionable when there is appreciable overlap of monolayer and multilayer adsorption.

In this work, we have used grand canonical Monte Carlo simulations to calculate nitrogen adsorption in several different classes of metal-organic frameworks. The BET surface areas are determined from the calculated isotherms and compared with available experimental results. We also calculated the accessible surface areas directly from the crystal structures. The accessible surface area is defined by the center of an adsorbate molecule rolling over the surface of the MOF. Comparing the accessible surface area with the BET area from the simulated isotherms allows for a critical test of the applicability of the BET method to these ultra-high surface area materials. In addition, we examine layering of the adsorbed molecules within the MOF to test the underlying assumptions of the BET method. The BET and accessible surface areas are examined with respect to MOF properties such as pore size, pore shape, and free volume. These results and recommendations for the most useful definition of MOF surface areas will be presented.