Mohammad Wahiduzzaman

• Physisorption of molecular hydrogen in nano- and micro-porous materials is a promising approach to improve the energy storage in mobile applications. Among all the absorbent materials that have been proposed for hydrogen adsorption, metal-organic frameworks (MOFs) and porous aromatic framework (PAFs) attracted significant attention because of their high surface areas and porosities. However, due to weak van der Waals interactions, physisorption of hydrogen in these adsorbents is only efficient at or near the temperature of liquid nitrogen. At low temperatures both free and adsorbed hydrogen in a confined pore possess significant quantum effects. State-of-the-art molecular simulation methods---grand canonical Monte Carlo (GCMC) with Feynman-Hibbs correction and quantized liquid density‑functional theory (QLDFT)---have been applied to investigate the role of quantum effects on the H2 adsorption process. Under explicit quantum mechanical treatment, simulations have been performed to explore the correlation between storage capacities and structural properties of porous solids. Adsorption calculations have been further extended to determine D2/H2 selectivity through quantum sieving in a narrow pore metal-organic framework. As required by the adsorption calculations, host-guest interaction potentials were parametrized and carefully validated. The accuracy and applicability of different sets of force fields have been discussed. Typical high-capacity H2 adsorbent materials have very large unit cells, in which the application of high-level quantum chemistry methods are prohibitively expensive. Some recent studies show that density-functional-based tight-binding (DFTB) method can be used to model large framework materials with sufficient accuracy and an affordable computational cost. Unfortunately, the method cannot be applied in a wide range of systems due to the limited availability of the parameters. In this work, a semi-automatic parametrization scheme for the electronic part of the DFTB parameters has been developed. Applying the scheme, a large part of the periodic table has been parametrized. The accuracy and transferability of the parameters have been tested for a wide variety of systems and found to give an excellent overall performance.