Development and application of computational quantum many-body methods for strongly correlated models and materials

Abstract

DFT, however, fails to explain the spectral properties of solids which have partially lled d and f shells. For this class of compounds, DFT quite often predicts a metallic ground state but experiments show that they are insulators. Due to the localized nature of d and f orbitals in space, screening of Coulomb interaction between electrons in those orbitals is poor. Thus, the electrons in d and f orbitals experience a much higher Coulomb repulsion than in the s/p orbitals, leading to strong correlation e ects, which in turn imply the break down of any effective one particle picture where the ground state wavefunction of the system is a combination of Slater determinants, and there are no well de ned one electron excitations in the system. Because of strong correlations, these systems exhibit interesting properties and phases. The materials which come under this category are termed as strongly correlated electronic systems (SCES). Typical examples of SCES[10] include cuprates, rare-earth compounds, actinides and transition metal oxides. Some of the features of strong correlation e ects include metal to Mott transition in V2O3[11{16], itinerant magnetism in transition metal oxides[17], giant magneto-resistance in manganites[18, 19], and hightemperature superconductivity in cuprates[20]. Theoretical studies of SCES require quantum many-body methods which are capable of handling strong correlations between electrons. Traditionally, these methods have been applied to studying model Hamiltonian's, that ignore material speci c information. With the advent of dynamical mean- eld theory (DMFT),