Computational Modelling of Perovskite-based Materials for Solid Oxide Fuel Cell Applications

Abstract

The development and use of clean and renewable modes of power generation is essential to address the increasing environmental and health concerns associated with fossil fuel consumption. Solid oxide fuel cells (SOFCs) are promising technologies for clean and efficient energy generation, as they can directly convert chemical energy into electrical power overcoming combustion limitations and utilising a variety of fuels. High temperatures are currently required for satisfactory ionic conduction at the electrolyte and efficient catalytic activity for the oxygen reduction reaction at the cathode, decreasing the lifetime and increasing the device’s cost. Hence, one of the main challenges associated with SOFC development is to decrease their operating temperatures to the intermediate temperature (IT) range (600-800 °C). Computational modelling is an essential tool in developing or improving IT-SOFC components, allowing researchers to predict, understand and explain the mechanisms behind experimentally observed properties. This thesis focus on the investigation of defect properties of perovskite-based materials and on the improvement of their ionic and/or electronic conduction properties, aiming their application as IT-SOFC components. LaGaO3-based materials have been suggested as suitable electrolytes, as they possess high conductivities at lower temperatures when compared to that of currently used yttria stabilized zirconia and samarium or gadolinium doped ceria. DFT calculations were performed to investigate the potential effects of a range of divalent dopants in the ionic conductivity and structure of LaGaO3 aiming to ascertain which dopants are best suited for solid electrolyte applications. Quantities such as doping and association energies were evaluated, to determine how easily the distinct dopants were accepted into the lattice and if there is any tendency of vacancy segregation around them, which could be detrimental to ionic transport. Sr(II), Ba(II) and Mg(II) were identified as the most suitable dopants, with steric and electronic defects playing a role in how easily these cations are accepted into the lattice. The local structural distortion introduced upon doping was observed to be an indicator of the tendency of vacancies to be trapped around dopants, possibly hindering ionic conductivity. The ionic conductivity of Sr and Mg doped LaGaO3 (LSGM) was then examined by means of molecular dynamics simulations, as the investigation of the diffusion and dynamics in doped LaGaO3 requires larger simulation cells. An interatomic potential was derived from ab initio data, as the currently available force fields are empirically derived and lack transferability, and has shown good agreement with experimental structural results. The conduction and structural properties of LSGM systems with a range of dopant concentrations (between 5 and 50 mol%) were then investigated. The calculated ionic conductivity and activation energy values were in good agreement with the available experimental data, with a total dopant concentration of around 20 mol% being concluded to be optimal. Local structure analysis was carried out to evaluate the tendency of vacancy clustering around dopants and vacancy ordering throughout the lattice, and the influence of dopant content in such effects. Mg-doping was observed to be more beneficial to ionic conduction properties, as vacancies tend to be trapped around Sr cations, deteriorating ionic transport. Vacancy ordering occurs independently of dopant identity, more intensely in systems with higher dopant content. Finally, the defect chemistry of La2NiO4, a layered Ruddlesden-Popper oxide and potential cathode material, was investigated. La2NiO4 is a mixed ionic and electronic conductor, with A- and B-site doping affecting the material’s transport properties in different ways. Usually, a trade-off between ionic and electronic conductivity improvements needs to be achieved. Density functional theory calculations were carried out to investigate the effects of a range of A- and B-site dopants in the material’s defect chemistry and conduction properties. Defect formation energies, transition level diagrams and preferable charge compensating mechanisms were investigated under distinct chemical environments so the optimal dopants for intermediate temperature cathode applications could be determined. The stability of oxygen defects was examined, as the Ruddlesden-Popper structure can accommodate both oxygen vacancies and interstitials; interstitials were predicted to be the dominant oxygen defect. The introduction of the selected A-site dopants is preferably charge compensated by the formation of electron holes, and hence could yield p-type electronic conductivity improvements. The calculated defect binding energies indicate that A-site dopants should not act as electron hole traps. The introduction B-site dopants, on the other hand, is preferably charge compensated with the introduction of oxygen interstitials, and hence could result in ionic conductivity improvements. Sr(II) and Co(III) were identified as the most suitable dopants for IT- SOFC applications.