Correlations between Atomic Structure and Dynamics in Porous Nano-domained Materials.

摘要

First principles methods are applied directly to explore the bulk properties of Solid Oxide Fuel Cell (SOFC) and Metal-Organic Framework (MOF) materials and to develop thermodynamic models via cluster expansions to explore the nanostructure and transport properties of SOFCs. Atom scaled ordering of dopants, defects, and magnetic moments within crystalline materials are of special interest in SOFC and lanthanum perovskite systems. A Genetic Algorithm optimizer for Cluster Expansion (CE) methods, a locally adaptive extension for CE methods, and a locally adaptive Lattice Monte Carlo algorithm are developed to enhance cluster expansion fit effectiveness, allowing study of previously prohibitively complex crystalline systems. Results of these methodological advancements resulted in studies used to suggest methods and means for experimental work to enhance properties in SOFC materials. Experimental models, pushing the bounds of the adaptive CE methods, are constructed for surface, magnetic, and interfaced systems. Studies of dopant/defect order in yttria-stabilized zirconia (YSZ) show a tendency for cations to aggregate, vacating significantly sized domains. This aggregation, along with the preferential association of oxygen vacancies with Y dopants, is likely important for the YSZ aging process in ionic conductivity. Studies of dopant/defect ordering in doped lanthanum chromate perovskties, (La1-xSrx)(Cr1-yFe y)O3-delta (LSCF) and (La1-xSrx)(Cr 1-yRuy)O3-delta (LSCR), suggest increased oxygen vacancy association with Ru atoms upon material reduction. Using the nudged elastic band (NEB) methodology, the large number of three-oxygen vacancy associates observed around reduced Ru dopants may provide a means for fast Ru diffusion via Cr-Ru swapping. No such oxygen vacancy association was seen in LSCF. Instead, long-range ordering of Fe with both Fe atoms and oxygen vacancies is observed, decreasing Fe's expected mobility. Utilizing the adaptive CE and locally adaptive Lattice Monte Carlo, differing mechanisms of diffusion degradation in YSZ were found to be likely, depending upon the operational redox environment of the material. In particular, increased trapping of oxygen vacancies by Y atoms is probable in the reducing environment found at the anode of the SOFC while increased Y aggregation is likely in the oxidative environment found at the cathode of the SOFC.