Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Band gaps of insulators from moment-functional-based spectral density functional theory

Frank Freimuth, Stefan Blügel, and Yuriy Mokrousov
Phys. Rev. B 108, 165137 – Published 19 October 2023

Abstract

Within the method of spectral moments it is possible to construct the spectral function of a many-electron system from the first 2P spectral moments (P=1,2,3,). The case P=1 corresponds to standard Kohn-Sham density functional theory (KS-DFT). Taking P>1 allows us to consider additional important properties of the uniform electron gas (UEG) in the construction of suitable moment potentials for moment-functional-based spectral density functional theory (MFbSDFT). For example, the quasiparticle renormalization factor Z, which is not explicitly considered in KS-DFT, can be included easily. In the four-pole approximation of the spectral function of the UEG (corresponding to P=4) we can reproduce the momentum distribution, the second spectral moment, and the charge response acceptably well, while a treatment of the UEG by KS-DFT reproduces from these properties only the charge response. For weakly and moderately correlated systems we can reproduce the most important aspects of the four-pole approximation by an optimized two-pole model, which leaves out the low-energy satellite band. From the optimized two-pole model we extract parameter-free universal moment potentials for MFbSDFT, which improve the description of the band gaps in Si, SiC, BN, MgO, CaO, and ZnO significantly.

  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
  • Figure
7 More
  • Received 9 June 2023
  • Accepted 6 October 2023

DOI:https://doi.org/10.1103/PhysRevB.108.165137

©2023 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. 108, Iss. 16 — 15 October 2023

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

Authorization Required


×

Images

  • Figure 1
    Figure 1

    Comparison of the momentum distribution function obtained within the four-pole model (4-Pole) to the one given in Refs. [26, 27] (OB) and to the one given in Ref. [4] (GGZ). The dimensionless density parameter is set to rs=3.

    Reuse & Permissions
  • Figure 2
    Figure 2

    Comparison between MkF(2+)=V(2+) as obtained from the four-pole model, the (n+1)-pole model of Sec. 2d with 81 poles, and the model of Ref. [5] discussed in Appendix pp1 (VZN). The vertical axis employs the unit of Ry2, where Ry=22maB2=13.6eV.

    Reuse & Permissions
  • Figure 3
    Figure 3

    Ratio of the charge response P0 and the density of states of the corresponding KS system at the Fermi energy D(EF) [Eq. (74)]. Four methods are compared. VWN: Eq. (77) using the parametrization of Vxc given in Ref. [1]. VDMC: Results from variational diagrammatic Monte Carlo as given in Ref. [29]. 4-Pole: Results for the four-pole model as computed from Eq. (79). 81-Pole: Results for the (n+1)-pole model as computed from Eq. (93).

    Reuse & Permissions
  • Figure 4
    Figure 4

    Comparison of the momentum distribution function obtained within the 41-pole model to the one given in Ref. [4] (GGZ). The dimensionless density parameter is set to rs=3.

    Reuse & Permissions
  • Figure 5
    Figure 5

    Quasiparticle renormalization factor Z as obtained from the model of Ref. [4] (solid line) and from the variational diagrammatic Monte Carlo calculations of Ref. [3] (circles). Dotted lines are linear interpolations and extrapolations of the variational diagrammatic Monte Carlo data.

    Reuse & Permissions
  • Figure 6
    Figure 6

    Density of states (DOS) of Si vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 7
    Figure 7

    Density of states (DOS) of SiC vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 8
    Figure 8

    Density of states (DOS) of diamond vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 9
    Figure 9

    Density of states (DOS) of BN vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 10
    Figure 10

    Density of states (DOS) of MgO vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 11
    Figure 11

    Density of states (DOS) of CaO vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 12
    Figure 12

    Density of states (DOS) of ZnO vs energy E as obtained in KS-DFT and in MFbSDFT. EF is the Fermi energy.

    Reuse & Permissions
  • Figure 13
    Figure 13

    Density of states (DOS) in the spin-unpolarized phase of Ni. KS-DFT and MFbSDFT (optimized two-pole model) are compared.

    Reuse & Permissions
  • Figure 14
    Figure 14

    Density of states (DOS) in the spin-unpolarized phase of Ni. KS-DFT and MFbSDFT (three-pole model) are compared.

    Reuse & Permissions
×

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
×