Dielectric embedding GW for weakly coupled molecule-metal interfaces

TL;DR

This paper introduces a computationally efficient GW-based dielectric embedding method for accurately determining electronic level alignments at weakly coupled molecule-metal interfaces, reducing costs while maintaining precision.

Contribution

The authors develop a dielectric embedding approach that enables GW calculations on large interfaces by focusing only on the molecular region, significantly lowering computational demands.

Findings

Successfully applied to benzene on Al(111) and graphite surfaces

Achieved accurate level alignment with reduced computational cost

Assessed GW approximations for weakly coupled interfaces

Abstract

Molecule-metal interfaces have a broad range of applications in nanoscale materials science. Accurate characterization of their electronic structures from first-principles is key in understanding material and device properties. The GW approach within many-body perturbation theory is state-of-the-art and can in principle yield accurate quasiparticle energy levels and interfacial level alignments that are in quantitative agreement with experiments. However, the interfaces are large heterogeneous systems that are currently challenging for first-principles GW calculations. In this work, we develop a GW-based dielectric embedding approach for molecule-metal interfaces, significantly reducing the computational cost of direct GW without sacrificing accuracy. To be specific, we perform explicit GW calculations only in the simulation cell of the molecular adsorbate, in which the dielectric effect of the metallic substrate is embedded. This is made possible via a real-space truncation of the substrate polarizability and the use of the interface plasma frequency in the adsorbate GW calculation. Here, we focus on the level alignment at weakly coupled molecule-metal interfaces, i.e., the energy difference between a molecular frontier orbital resonance and the substrate Fermi level. We demonstrate our method and assess a few GW-based approximations using two well-studied systems, benzene adsorbed on the Al (111) and on the graphite (0001) surfaces.