Spectral properties from an efficient analytical representation of the G ​ W GW self-energy within a multipole approximation

Dario A. Leon, Kristian Berland, Claudia Cardoso · 2025

Read Spectral properties from an efficient analytical representation of the G ​ W GW self-energy within a multipole approximation by Dario A. Leon, Kristian Berland, Claudia Cardoso (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-09121.txt.

Introduction

Dario Leon, Kristian Berland, and Claudia Cardoso give a fast math short cut for a key term in electron theory, in a 2025 paper. The term, called self energy, shapes how light and charge marks of matter are found. Full sums for it cost much time, most of all across wide energy bands.

Their short cut fits the term from a few points in a math plane with a pole based form. Picking the fit points with care keeps speed and truth both high. Tests on solids, thin sheets, and small groups of atoms check out well. The method also gives a clean way to set a key weight factor.

Catalog record

CREATORDario A. Leon, Kristian Berland, Claudia Cardoso
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-09121

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Spectral properties from an efficient analytical representation of the G ​ W GW self-energy within a multipole approximation Dario A. Leon, Kristian Berland, Claudia Cardoso Abstract We propose an efficient analytical representation of the frequency-dependent G ​ W GW self-energy Σ \Sigma via a multipole approximation (MPA- Σ \Sigma ). The MPA self-energy model is interpolated from a small set of numerical evaluations of Σ \Sigma in the complex frequency plane, similar to the MPA interpolation developed for the screened Coulomb interaction (MPA- W W ) [D. A. Leon et al. , Phys. Rev. B 104 , 115157 (2021)]. Crucially, MPA- Σ \Sigma enables a multipole representation for the interacting Green’s function G G (MPA- G G ), and in turn, access to all the spectral properties, including quasiparticle energies (QP) and renormalization factors beyond the linearized QP equation. We validate the MPA- Σ \Sigma and MPA- G G approaches for a diverse set of systems: bulk Si, Na and Cu, monolayer MoS 2 , the NaCl ion-pair, and the F 2 molecule. We show that, just as for MPA- W W , an appropriate choice of frequency sampling in MPA- Σ \Sigma is critical to guarantee computational efficiency and high accuracy. Moreover, the combined MPA- W W and MPA- Σ \Sigma scheme considerably reduces the cost of full-frequency self-energy calculations, especially for spectral band structures over a wide energy range. I Introduction In condensed matter physics, first principle methods such as density functional theory (DFT) in the Kohn-Sham (KS) approximation provide accurate ground-state properties and have been immensely useful for understanding the electronic structure of materials. However, they fail to reliably provide accurate band structures, which requires including many-body effects beyond the mean field DFT level. The description of electron addition or removal energies and related excited-state properties is usually treated with methods such as the GW approximation, based on the Green’s function formalism Aryasetiawan and Gunnarsson 1998 ; Hedin 1999 ; Onida et al. 2002 ; Martin et al. 2016 ; Reining 2018 ; Marzari et al. 2021 . In common GW implementations, the Green’s function G and the screened Coulomb potential W are constructed perturbatively. Starting from DFT, the KS quasiparticle (QP) energies are corrected by an exchange-correlation self-energy \Sigma . This correction can be obtained iteratively within a self-consistent GW approach, or done in a computationally cheaper one-shot G_{0}W_{0} procedure. Since the imaginary part of G is closely related to the spectral function obtained from photoemission experiments Damascelli 2004 ; Hüfner 2013 , a dynamical self-energy \Sigma(\omega) can account for many-body features, such as finite QP lifetimes and satellite structures Strinati et al. 1980 ; Strinati et al. 1982 ; Hedin 1999 ; Dolado et al. 2001 ; Marini et al. 2002a ; Kheifets et al. 2003 ; Arnaud et al. 2005 ; Cazzaniga 2012 ; Reining 2018 ; Zhou et al. 2020 . G_{0}W_{0} is the state-of-the-art ab initio method for the description of angle-resolved photoemission and inverse photoemission spectroscopy measurements, giving generally a very accurate agreement with experiment (see, e.g., Refs. Hybertsen and Louie 1985 ; van Schilfgaarde et al. 2006 ; Hüser et al. 2013 ; Golze et al. 2019 ). More accurate spectral functions can be obtained with self-consistent approaches, including cumulant expansions of \Sigma and vertex corrections (see, e.g., Refs. Kheifets et al. 2003 ; Martin et al. 2016 ; Gumhalter et al. 2016 ; Zhou et al. 2018 ; Nery et al. 2018 ; Reining 2018 ; Zhou et al. 2020 ). The GW self-energy \Sigma(\omega) is given by a frequency convolution of G(\omega) and W(\omega) . This convolution can be evaluated with different full-frequency (FF) methods, based on numerical integrations along the frequency real axis Marini et al. 2002b ; Shishkin and Kresse 2006 ; Hüser et al. 2013 ; Liu et al. 2015 , or through an integration in the complex frequency plane using contour deformation and analytic continuation techniques Godby et al. 1988 ; F. Aryasetiawan in 2000 ; Kotani et al. 2007 ; Daling et al. 1991 ; Engel et al. 1991 ; Duchemin and Blase 2020 . Such numerical FF evaluations tend to be computationally expensive. A less costly alternative is to represent W (or the dielectric function) with a simple model, such as in the plasmon pole approximation (PPA), that allows for an analytical integration of the frequency convolution in \Sigma Hybertsen and Louie 1986 ; Zhang et al. 1989 ; Godby and Needs 1989 ; von der Linden and Horsch 1988 ; Engel and Farid 1993 , but in many cases this has limited accuracy. Higher accuracy can be obtained with multipole models and Padé approximants Farid et al. 1991 ; Engel et al. 1991 ; Lee and Chang 1994 ; Soininen et al. 2003 , including the recently developed MPA- W method Leon et al. 2021 ; Leon et al. 2023 ; Guandalini et al. 2024 . Analytical models of the dielectric response are also widely used in the study of optical and electronic properties of materials Raether 1980 ; Giuliani and Vignale 2005 ; Pines 2018 . They have been used in the study of, e.g., optical excitations Allen and Mikkelsen 1977 ; Smith and Segall 1986 , electron energy loss Lee and Chang 1996 ; Jin and Chang 1999 , and x-ray absorption spectra Kas et al. 2007 ; Kas et al. 2009 . Simple models have also been used to account for dynamical effects arising from electron-hole interactions in doped systems Liang and Yang 2015 ; Champagne et al. 2023 , or in the ab initio description of plasmon-phonon hybridization in doped semiconductors Lihm and Park 2024 . Less common is the use of models for \Sigma Riegera et al. 1999 ; Soininen et al. 2005 ; van Setten et al. 2015 ; Chiarotti et al. 2022 , the interacting G Chiarotti et al. 2022 ; Chiarotti et al. 2024 ; Ferretti et al. 2024 ; Quinzi et al. 2025 , or for total energy calculations Ismail-Beigi 2010 ; Chiarotti et al. 2022 ; Chiarotti et al. 2024 ; Ferretti et al. 2024 ; Quinzi et al. 2025 ; Guo and Liu 2024 , and usually these approaches mainly aim at improving the computational efficiency of such calculations. In this work, we present an efficient multipole approximation for the self-energy (MPA- \Sigma ). This method yields simple analytical representations of all GW operators, including a multipole-Padé representation of the Green’s function (MPA- G ). Moreover, the combination of MPA- \Sigma with the previous MPA- W method considerably reduces the cost of evaluating \Sigma and G in its full-frequency domain. Although this study is limited to the G_{0}W_{0} approximation, extending it to higher levels of theory, e.g., self-consistent GW , and the inclusion of vertex corrections or cumulant expansions, is straightforward. The paper is organized as follows: In the methods section (Sec. II ) we summarize the general GW equations (Sec. II.1 ), and the previously introduced MPA- W method (Sec. II.2 ). Thereafter, we present the new MPA- \Sigma (Sec. II.3 ) and MPA- G (Sec. II.4 ) approaches, and provide computational details of the GW calculations (Sec. II.5 ). In the results section (Sec. III ) we benchmark

Read the complete text