384584 Combining Ab Initio and Classical Simulation Techniques in a High-Throughput Fashion to Assess Hydrogen Uptake in Metal-Organic Frameworks
New high-throughput techniques, both experimental and computational, are beginning to have a significant impact on materials research and design. One class of materials impacted by this new trend is metal-organic frameworks (MOFs). MOFs are self-assembled from inorganic nodes and organic edges or linkers, resulting in porous, crystalline materials. MOF building blocks (nodes and linkers) can be interchanged and/or altered, resulting in a virtually limitless number of possible structures. Particularly exciting about these materials is the ability to tailor them, through judicious choice of building blocks, for particular applications. Additionally, relevant chemical functionalities such as hydroxyl, amino, and metal alkoxides, may be introduced, therefore adding to the chemical diversity of these materials.
We used novel computational high-throughput techniques to address the hydrogen storage challenge using MOFs. Hydrogen storage remains as, arguably, the biggest challenge towards the development of future fuel cell vehicles. The United States Department of Energy (DOE) has set hydrogen storage targets of 7.5 wt.% and 70 g/L at a minimum temperature of 243 K and a maximum pressure of 100 bar; to date, no material has met these targets. Thanks to their large pore volumes and surface areas, MOFs excel at hydrogen storage near cryogenic temperatures. However, they falter near room temperature due to the low heats of adsorption (~10 kJ/mol) of the hydrogen molecules in MOFs. In order to counteract this, we introduced open metal sites in the form of metal alkoxides. We used quantum mechanical calculations to calculate hydrogen binding energies on a series of metals, and we fit a Morse potential to the interactions to use as a force field in our grand canonical Monte Carlo (GCMC) simulations. We determined Mg alkoxide to be a promising candidate for improving hydrogen storage in MOFs.
Then, we modified our recently published methodology for computationally building MOFs to include the Mg alkoxide groups. Briefly, we took experimentally synthesized structures and disassembled them into their building blocks (inorganic nodes and organic linkers). We then introduced the Mg alkoxide functionalities into the linkers and recombined them with the nodes to form the MOF structures. This resulted in over 18,000 structures with varying Mg alkoxide content, as well as a wide range of physical properties such as void fraction, pore volume, and surface area. We used GCMC simulations to calculate hydrogen uptake in all generated structures at 2 bar and 100 bar at 243 K. The most promising structure in gravimetric storage is predicted to adsorb 9.35 wt.% hydrogen, meeting the DOE target. Volumetrically, the most promising structure is predicted to adsorb 51 g/L of hydrogen, falling just short of the DOE target. We also found trade-offs between gravimetric and volumetric uptake along with new structure-property relationships. Finally, our computational screening techniques allowed us to assess what could be the performance limits of MOFs for hydrogen storage applications near room temperature.