First Principles Study of Double Perovskites and Group III-V Compounds

The broad aim of this dissertation is to develop an atomic scale understanding of the magnetic and electronic properties of double perovskites, a class of materials that hold a lot of promise to realize new multifunctional devices. We have used first-principles density functional theory (DFT) calculations to develop this understanding.;In the first section, we focus on double perovskites with half metallic properties. We begin with the most widely studied half metallic double perovskite, Sr2FeMoO6. Even after more than a decade of extensive research, it has not yet been possible to realize the high degree of spin-polarization that has been theoretically predicted in Sr2FeMoO6. We find point defects to be playing a major role in degrading its half metallic properties. We determine transition-metal antisites and oxygen vacancies to be the thermodynamically stable defects, and predict Mo-rich Sr2FeMoO 6 and stoichiometric Sr2FeMoO6 with antisite disorder to have poor spin-polarization.;We have then used our understanding of the magnetic interactions that result due to antisites disorder in Sr2FeMoO6 to predict Ca2MnRuO6 as a material which should allow high levels of spin-polarized conduction, even in a complete disordered form.;Next, we have studied the magnetic interactions in a recently discovered double perovskite Sr2CoOsO6, where the two magnetic ions Co and Os order independently of each other, a behavior that has never been observed in any other magnetic material. By calculating the strength of exchange interactions between different pairs of magnetic ions in the compound, we attribute the observed magnetic behavior to the weak nearest neighbor Co--O--Os interactions compared to the stronger Co--O--Os--O--Co and Os--O--Co--O--Os interactions. We explain this weak Co--O--Os interaction to be due to the small overlap of the Co 3d and Os 5d orbitals.;In the next section, we apply the method of determining the chemical potential of the constituent elements in multicomponent systems, which we had developed to calculate the stable defects in Sr2FeMoO 6, to a simpler binary semiconductor InP and expand it into a more rigorous and general theory, which we found had been missing in the literature when we examined the point defects in Sr2FeMoO6. After identifying the stable neutral defects in both stoichiometric and non-stoichiometric InP, we determine their defect level within the experimental band gap, using a combination of computationally fast traditional exchange--correlation functionals used in DFT, which underestimate the experimental band gap, and computationally expensive but more accurate hybrid functionals. We combine the neutral-defect formation energies with the position of the defect states and suggest a method to calculate both, chemical potentials of the constituent elements, and the formation of energies of charged defects, as a function of the Fermi level.;In the final section, we suggest a method to study the electronic properties of interfaces, which is general enough to address the common questions at interfaces between complex materials such as e.g. found in semiconductor systems or complex oxide heterostructures. Using wurtzite AlN/GaN interface as a test system, we calculate the band line-up and interface charge density due to spontaneous and piezoelectric polarization, using a combination of first principles real-space band structure and electrostatics. Using this method, we show the interface charge density in the AlN/GaN system to be independent of the layer thicknesses of the two materials, with their value being identical to the interface charge density calculated using bulk constants.