Qin, Tian2019-08-202019-08-202019-05https://hdl.handle.net/11299/206406University of Minnesota Ph.D. dissertation.May 2019. Major: Earth Sciences. Advisor: Renata Wentzcovitch. 1 computer file (PDF); xiv, 113 pages.This thesis presents a series of ab initio studies on two major topics of interest in geochemistry: a) water speciation in forsterite, and b) iron and silicon isotope fractionation among bridgmanite and metallic phases. The first topic relates to the deep water cycle in the Earth; we studied the stability of two charge-balanced hydrous defects, (2H)_X^Mg and (4H)_X^Si , in the forsterite. We systematically searched the possible configurations of these two defects and then included the contribution of vibrational energy and configurational entropy in the calculation of the formation energies of both defects. Our results reveal that the configurational entropy plays a key role in stabilizing the (2H)_X^Mg defects, and further, the water speciation in forsterite is shown to be influenced by temperature and pressure. A Python program called qha" has been developed to calculate the thermodynamic properties of these multi-configurational hydrous defects. The second topic addresses the enigmatic and paradoxical Fe isotope composition found in the terrestrial basalts and the Si isotope composition of the bulk silicate Earth (BSE). The Fe and Si isotope fractionation factor between the mantle and core phases have been calculated. The mineral phases studied here include Fe-bearing bridgmanite, pure HCP Fe and an Fe-Si alloy. The different valence, structural site, spin states of Fe in bridgmanite are taken into consideration in the calculation of Fe isotope fractionation between silicate and metallic core. We show that the low spin Fe in the bridgmanite has a strong preference to enrich heavy isotopes. The simple-mass-balance calculations suggest that the core-mantle segregation can lead to the mantle being heavier in Fe isotopes compared to chondrites. The calculated Si isotope fractionation factor between silicate phase and the Fe-Si alloy suggests that the silicate phase enrich heavy Si isotopes. However, the calculated fractionation factor is much smaller than previous experimental estimation; one of the consequences is that it is difficult to match the silicon content in the core. Thus, our study suggests that the Earth might be non-chondritic in Si isotope or that calculations for the melt phases are really required for this case. While these big questions have not been resolved, we shed light on a possible origin for the anomalous Fe isotope compositions of basalts, and the possible Si isotope composition of the BE and BSE, and we have successfully explained some apparently controversial facts concerning hydrous defects in forsterite and clearly established a path for addressing more realistically the relative stability of hydrous defects.enThesis or Dissertation