Accurate treatment of metallic screening in Many-Body calculations from first principle

Many-Body Perturbation Theory calculations within the GW approximation have proven to be a reliable scheme for the computation of quasiparticle (QP) band structures of materials. While QP corrections in metallic systems are generally smaller than in semiconductors, they are crucial for accurate spectroscopic descriptions, especially in low-dimensional systems.

The computational cost of GW poses however challenges, particularly in accurately describing the screened potential in both frequency and k-space for metals. One of the challenges is the inclusion of the long-wavelength limit of intraband transitions, usually treated by adding a Drude term at the plasmon energy. However, the ab-initio determination of the plasmon energy is generally demanding. In addition, the Drude method may be too simplistic for certain metals and semimetals, since more than one pole may be needed in the description of interband transitions1.

In this colloquium, I discuss a recently developed method for performing efficient QP computations in metals. This approach, initially developed for 2D semiconductors, involves Monte Carlo integration and interpolation techniques2,3, leading to significant speed-ups in k-point convergence4. Extending its applicability to 3D and 2D metals, I demonstrate the accurate evaluation of the screened potential, allowing for efficient QP computations without additional parameters. The method reproduces QP band structures with reduced k-point grids, offering a streamlined approach for studying metallic systems.

1) D. A. Leon, C. Cardoso, T. Chiarotti, D. Varsano, E. Molinari, and A. Ferretti. Phys. Rev. B, 104, 115157 (2021).

2) F. H. da Jornada, D. Y. Qiu, and S. G. Louie. Phys. Rev. B 95, 035109 (2017).

3) W. Xia , W. Gao, G. Lopez-Candales, Y. Wu, W. Ren, W. Zhang, and P. Zhang. npj Comput. Mater. 6, 118 (2020).

4) A. Guandalini, P. D’Amico, A. Ferretti, and D. Varsano. npj Computational Materials 9, 44 (2023).

Seminar realized in the framework of the funded project:
-PRIN2017 “Excitonic insulator in two-dimensional long-range interacting systems” (EXC-INS)
-National Centre for HPC, Big Data and Quantum Computing
-MaX - Materials design at the eXascale- GA No. 101093374