Computational Modeling of Gas Adsorption, Separation, and Reactivity within Coordinatively Unsaturated Metal-Organic Framework Materials

Abstract

In this work, computational methodologies are used to investigate the behavior of Metal-Organic Frameworks (MOFs) that are of potential utility in gas separation ap- plications. MOFs are three-dimensional porous materials that are desirable due to their high porosity and internal surface area. Given the myriad of possible framework topologies, computational tools are necessary in order to aid and complement experimental efforts. The focus of this dissertation is specifically coordinatively unsaturated MOFs, a subclass of MOFs for which there are additional computational challenges in treating the exposed metal site. The collection of studies presented here are grouped into four categories: force field parameterization, gas reactivity within MOFs, nature of adsorbate-MOF bonding, and multireference treatment of metal-metal bonds. Ab initio force fields are parameterized for various MOF-gas interactions, where the the NonEmpirical Modeling (NEMO) philosophy is adopted. Wave-Function Theory (WFT) is used for the calculation of the reference energies of the intermolecular terms. Moller-Plesset Perturbation Theory to 2nd order (MP2) and Complete-Active Space Perturbation Theory to 2nd order (CASPT2) are applied for closed-shell and open-shell cases, respectively. These derived force fields are utilized in Grand Canonical Monte Carlo (GCMC) simulations for the MOF + adsorbate systems. Results from GCMC simulations include adsorption isotherms, Henry coefficients, and isosteric heats of adsorption that are compared with the available experimental data, demonstrating the predictive capabilities of this computational procedure. A reaction mechanism between CO2 and amines grafted within the pores of a MOF is proposed based on DFT results, and both DFT and CASPT2 results are utilized to elucidate the nature of a reactive Fe-Oxo intermediate at the exposed Fe site in the MOF. Coupled Cluster (CC), MP2, CASPT2, and DFT are all used to rationalize adsorbate-MOF bonding trends, and the Extended-Transition State Natural Orbitals for Chemical Valence (ETS-NOCV) is used as a comparative tool. Complete-Active Space Self-Consistent Field (CASSCF) and CASPT2 are used for the studies of metal-metal multiple bonded species, and compared with results from DFT.