Construction of the spectral function from noncommuting spectral moment matrices

Frank Freimuth, Stefan Blügel, and Yuriy Mokrousov
Phys. Rev. B 106, 045135 – Published 25 July 2022

Abstract

The LDA+U method is widely used to study the properties of realistic solids with strong electron correlations. One of its main shortcomings is that it does not provide direct access to the temperature dependence of material properties such as the magnetization, the magnetic anisotropy energy, the Dzyaloshinskii-Moriya interaction, the anomalous Hall conductivity, and the spin-orbit torque. While the method of spectral moments allows us in principle to compute these quantities directly at finite temperatures, the standard two-pole approximation can be applied only to Hamiltonians that are effectively of single-band type. We do a first step to explore if the method of spectral moments may replace the LDA+U method in first-principles calculations of correlated solids with many bands in cases where the direct assessment of the temperature dependence of equilibrium and response functions is desired: The spectral moments of many-band Hamiltonians of correlated electrons do not commute and therefore they do not possess a system of common eigenvectors. We show that nevertheless the spectral function may be constructed from the spectral moments by solving a system of coupled nonlinear equations. Additionally, we show how to compute the anomalous Hall conductivity of correlated electrons from this spectral function. We demonstrate the method for the Hubbard-Rashba model, where the standard two-pole approximation cannot be applied because spin-orbit interaction (SOI) couples the spin-up and the -down bands. In the quest for new quantum states that arise from the combination of SOI and correlation effects, the Hartree-Fock approximation is frequently used to obtain a first approximation for the phase diagram. We propose that using the many-band generalization of the self-consistent moment method instead of Hartree-Fock in such exploratory model calculations may improve the accuracy significantly, while keeping the computational burden low.

  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Received 27 May 2022
  • Revised 11 July 2022
  • Accepted 14 July 2022

DOI:https://doi.org/10.1103/PhysRevB.106.045135

©2022 American Physical Society

Physics Subject Headings (PhySH)

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

Frank Freimuth1,2,*, Stefan Blügel1, and Yuriy Mokrousov1,2

  • 1Peter Grünberg Institut and Institute for Advanced Simulation, Forschungszentrum Jülich and JARA, 52425 Jülich, Germany
  • 2Institute of Physics, Johannes Gutenberg University Mainz, 55099 Mainz, Germany

  • *Corresponding author: f.freimuth@fz-juelich.de

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand
Issue

Vol. 106, Iss. 4 — 15 July 2022

Reuse & Permissions
Access Options
Author publication services for translation and copyediting assistance advertisement

Authorization Required


×
×

Images

×

Sign up to receive regular email alerts from Physical Review B

Log In

Cancel
×

Search


Article Lookup

Paste a citation or DOI

Enter a citation
×