Bands, resonances, edge singularities and excitons in core level spectroscopy investigated within the dynamical mean-field theory

Abstract

PACS 78.70.Dm -X-ray absorption spectra PACS 79.60.-i -Photoemission and photoelectron spectra PACS 78.20.Bh -Theory, models, and numerical simulation Abstract -Using a recently developed impurity solver we exemplify how dynamical mean field theory captures band excitations, resonances, edge singularities and excitons in core level x-ray absorption (XAS) and core level photo electron spectroscopy (cPES) on metals, correlated metals and Mott insulators. Comparing XAS at different values of the core-valence interaction shows how the quasiparticle peak in the absence of core-valence interactions evolves into a resonance of similar shape, but different origin. Whereas XAS is rather insensitive to the metal insulator transition, cPES can be used, due to nonlocal screening, to measure the amount of local charge fluctuation.Core level photoemission (cPES) and core level x-ray absorption spectroscopy (XAS) have long been valuable tools in the field of material research for a huge range of compounds with a different degree of correlations [1]. For example, XAS can probe the unoccupied density of states in GaAs, Al or Hydrocarbons [2], local properties of correlated d-shells in transition-metal compounds like the cuprates [3][4][5], or the atomic like ground state symmetry of rare earth ions in heavy Fermion compounds and impurities in an aluminum garnet used as laser medium [6][7][8][9]. Interestingly, the same experiment seems to measure a different observable (empty density of states or local symmetry of the occupied wave-function) depending on the amount of correlations in the material. This dichotomy is also present in theory. The theoretical efforts for the description of cPES and XAS can roughly be divided into two approaches based on an itinerant or local starting point.In the itinerant approach one approximates the interactions between electrons by a (mean-field) potential. As a result one obtains a set of freely moving particles. This is the case for Hartee-Fock or density functional theory (in the local density approximation) calculations. On this level of theory XAS is identified as the unoccupied single particle density of states and cPES as a delta-function representing the occupied core density of states [10]. The core-valence interaction can be modeled as an additional potential that suddenly arrises after the photon absorption, leading to an edge singularity in the spectral function [11][12][13][14]. For many systems, including most of the transition-metal and rare earth oxides, it has been realized early on that the inclusion of the explicit core-valence interactions beyond a mean-field potential is crucial. Many of these core level edges are much more determined by the local multiplet structure than by the band-structure of the material. Sawatzky and coworkers used local models approximating a solid by a single atom in an effective potential (crystal field) or by a small cluster (ligand field theory) [1,7]. These cluster models can be extended to include the band width of the material at the level of an...

Direct Observation of Cr3+ 3d States in Ruby: Toward Experimental Mechanistic Evidence of Metal Chemistry

“…3,9,27 The Cr 3+ electronic configurations of relevance in our description are 3d 3 and 2p 5 3d 4 . They result in atomic multiplets, which are described by 3d–3d Coulomb and 2p–3d Coulomb and exchange interactions parametrized in Slater–Condon integrals F dd k , F pd k (Coulomb) and G pd k (exchange) for Hartree–Fock calculations.…”

Realistic theory of electronic correlations in nanoscopic systems

“…This is not the only advance in numerical algorithms for strongly correlated electron systems that we discuss in this review paper. As described in Section 3.2, we implemented a variational version of exact diagonalization [32] as well as the computation of core level X-ray absorption and core level photo electron spectroscopy [117] in the framework of a new-generation Hamiltonian-based solver for DMFT [30]. This is then extended to the case of multiorbital physics in nanostructures, which we approach with our combined CT-QMC and ED impurity solvers in order to establishing links from realistic structures to simple models.…”

Natural-orbital impurity solver and projection approach for Green's functions

“…The same is true for observables that only need a limited resolution, e.g. several forms of core-level spectroscopy where the fine details of the spectra are smeared out by the large core-hole lifetime [5,49]. For general materials, one often needs to correctly describe states with a bandwidth on the order of a Rydberg with a resolution better than the smallest energy scale (such as the crystal fields as small as tens of meV in some rare-earth compounds).…”

“…They describe a wide range of quantum mechanical problems that involve a subsystem with a limited number of degrees of freedom (an impurity) coupled to a much larger system (a bath) that contains a quasi-continuum of degrees of freedom. Examples include the Kondo and heavy-fermion systems [1][2][3], core-level X-ray spectroscopy [4,5], tunneling in dissipative systems [6], and various problems in quantum optics [7]. In the past years, the interest in impurity models has also been reinvigorated by the continuous development of dynamical mean-field theory (DMFT) [8,9], in which the correlated lattice problem is mapped self-consistently to an effective impurity model.…”