Ground- and excited-state properties of inorganic solids from full-potential density-functional calculations

The development in theoretical condensed-matter science based on density-functional theory (DFT) has reached a level where it is possible, from “parameter-free” quantum mechanical calculations to obtain total energies, forces, vibrational frequencies, magnetic moments, mechanical and optical properties and so forth. The calculation of such properties are important in the analyses of experimental data and they can be predicted with a precision that is sufficient for comparison with experiments. It is almost impossible to do justice to all developments achieved by DFT because of its rapid growth. Hence, it has here been focused on a few advances, primarily from our laboratory. Unusual bonding behaviors in complex materials are conveniently explored using the combination of charge density, charge transfer, and electron-localization function along with crystal-orbital Hamilton-population analyses. It is indicated that the elastic properties of materials can reliably be predicted from DFT calculations if one takes into account the structural relaxations along with gradient corrections in the calculations. Experimental techniques have their limitations in studies of the structural stability and pressure-induced structural transitions in hydride materials whereas the present theoretical approach can be applied to reliably predict properties under extreme pressures. From the spin-polarized, relativistic full-potential calculations one can study novel materials such as ruthenates, quasi-one-dimensional oxides, and spin-, charge-, and orbital-ordering in magnetic perovskite-like oxides. The importance of orbital-polarization correction to the DFT to predict the magnetic anisotropy in transition-metal compounds and magnetic moments in lanthanides and actinides are emphasized. Apart from the full-potential treatment, proper magnetic ordering as well as structural distortions have to be taken into account to predict correctly the insulating behavior of transition-metal oxides. The computational variants LDA and GGA fail to predict insulating behavior of Mott insulators whereas electronic structures can be described correctly when correlation effects are taken into account through LDA+U or similar approaches to explain their electronic structures correctly. Excited-state properties such as linear optical properties, magneto-optical properties, XANES, XPS, UPS, BIS, and Raman spectra can be obtained from accurate DFT calculations.