Abstract
Molecular clusters representative of protonated, neutral, and deprotonated sites on a forsterite (Mg2SiO4) surface were employed to facilitate examination of Mg–Obr bond-breaking via DFT calculations with the B3LYP/6-31+G(d,p) methodology. Hydrolysis reactions of the molecular clusters with a H2O molecule yielded barrier heights of 21, 54, and 39 kJ/mol for protonated, neutral, and deprotonated sites in the gas-phase, respectively, and the rate constants calculated using these barrier heights were 5.7x108, 2.7x104, and 2.2x106 s-1, respectively. Aqueous-phase calculations on the gas-phase structures were also performed, and the barrier heights were 33, 40, and 21 kJ/mol for the protonated, neutral, and deprotonated models. Rate constants were 4.3x106, 6.1x105, and 6.0x108 s-1. For models energy-minimized in the aqueous-phase the barrier heights were 42, 44, and 40 kJ/mol, and the rate constants were 1.4x107, 3.0x104, and 9.9x105 s-1, respectively. These differences highlight the importance of modeling structures with inclusion of solvent effects. Rates of Mg2+ release from the forsterite surface were predicted using these rate constants and models of the reactive site density and the H+ or OH– surface speciation. These calculations are consistent with a more rapid rate of Mg2+ release under acidic conditions even though the activation energy barriers are equivalent within computational uncertainty. A comparison of these results to previous data shows that the predicted rates are much faster than experimentally measured dissolution rates, and this suggests that the Mg–Obr bond break is a rapid process which is a component of Mg2+ release from the surface consistent with previous experimental observation of preferential Mg2+ leaching from forsterite. A dissolution mechanism involving polymerization and hydrolysis of Si–Obr–Si linkages is discussed that is consistent with the discrepancy between Mg2+ release rates and dissolution rates of forsterite.