A first principles study of zeolites and related aluminosilicates

First principles molecular dynamics techniques were used to study aluminosilicate structures in order to better understand the chemistry of these materials. Both structural and chemical aspects of aluminosilicate bonding were studied in two systems of central importance to science and engineering.; First we show that the zero pressure phase of Al₂O₃ (corundum) undergoes as series of pressure induced phase transformations at 78 GPa and 223 GPa. The first transformation occurs at pressures commonly obtainable in diamond anvil apparatuses—suggesting important consequences for the ruby pressure scale. The second phase transformation occurs at pressures found in the lower mantle where Al₂O₃ is suspected of residing. This transformation could then have ramifications to the corresponding phase behavior of lower mantle minerals.; We next studied the chemistry of aluminum substitution in sodalite (a prototypical zeolite). We show that the so-called “framework collapse” (tetrahedral rotation) of sodalite is a direct result of Al substitution and has little to do with framework/ion interactions as first suggested by Pauling and others. This is a clear case where cluster calculations fail to describe physical bonding properties of aluminosilicate materials and fully periodic calculations are necessary.; We also offer evidence that the frontier states (at the valence band edge) at least partially dictate the location of extra-framework ions. These states are suspected of being responsible for the negative charged regions of the sodalite framework that interact Coulombicly with Na+ ions. This rationalization explains, for instance, the apparent bonding of H+ ions to oxygens neighboring framework aluminum atoms. It is also consistent with reaction theory which suggest such charge transfer processes are largely accomplished via the frontier states.; We also show that the static lattice approximation is valid in molecular dynamics simulations of methane diffusion in AlPO₄-5 (a straight pore zeolite). Diffusivities calculated using classical pairwise additive interaction potentials in both a rigid and fully flexible lattice where found to be nearly identical. This validation should prove useful in future first principles treatments of diffusion in straight pore zeolites, where a dynamic lattice model would be prohibitively time consuming.