First-principles treatment of Mott insulators: linearized QSGW+DMFT approach

Sangkook Choi, Audrey Kutepov, Kristjan Haule, Mark Van Schilfgaarde, Gabriel Kotliar

Research output: Contribution to journalArticlepeer-review

50 Citations (Scopus)
163 Downloads (Pure)


The theoretical understanding of emergent phenomena in quantum materials is one of the greatest challenges in condensed matter physics. In contrast to simple materials such as noble metals and semiconductors, macroscopic properties of quantum materials cannot be predicted by the properties of individual electrons. One of the examples of scientific importance is strongly correlated electron system. Neither localized nor itinerant behaviors of electrons in partially filled 3d, 4f, and 5f orbitals give rise to rich physics such as Mott insulators, high-temperature superconductors, and superior thermoelectricity, but hinder quantitative understanding of low-lying excitation spectrum. Here we present a new first-principles approach to strongly correlated solids. It is based on a combination of the quasiparticle self-consistent GW approximation and the dynamical mean-field theory. The sole input in this method is the projector to the set of correlated orbitals for which all local Feynman graphs are being evaluated. For that purpose, we choose very localized quasiatomic orbitals spanning large energy window, which contains most strongly hybridized bands, as well as upper and lower Hubbard bands. The self-consistency is carried out on the Matsubara axis. This method enables the first-principles study of Mott insulators in both their paramagnetic and antiferromagnetic phases. We illustrate the method on the archetypical charge transfer correlated insulators La2CuO4 and NiO, and obtain spectral properties and magnetic moments in good agreement with experiments.
Original languageEnglish
Article number16001
JournalNPJ Quantum Materials
Publication statusPublished - 27 Jul 2016


Dive into the research topics of 'First-principles treatment of Mott insulators: linearized QSGW+DMFT approach'. Together they form a unique fingerprint.

Cite this