Low-Scaling GW Algorithm Applied to Twisted Transition-Metal Dichalcogenide Heterobilayers

Area of Science:

• Computational materials science
• Quantum chemistry
• Condensed matter physics

Background:

• The GW method is crucial for calculating electronic band structures.
• High computational costs limit the application of traditional GW algorithms.
• Efficient methods are needed for studying complex materials.

Purpose of the Study:

• To present a novel, low-scaling, and efficient GW algorithm.
• To overcome the computational limitations of existing GW methods.
• To enable large-scale electronic structure calculations.

Main Methods:

• Development of a periodic, low-scaling GW algorithm.
• Leveraging the locality of Gaussian basis sets and polarizability.
• Implementation for G0W0 calculations on large material systems.

Main Results:

• Achieved a significant speed-up (4 orders of magnitude) compared to plane-wave methods.
• Successfully performed G0W0 calculations on a large MoSe2/WS2 bilayer (984 atoms).
• Demonstrated unprecedented computational efficiency for electronic excitation studies.

Conclusions:

• The new algorithm dramatically reduces computational cost for GW calculations.
• Enables large-scale electronic structure studies previously intractable.
• Opens new avenues for nanoscale electronic excitation research.