An ab initio approach to energy alignment and charge-state prediction of adsorbates on ultrathin insulators

Abstract

The rapid progress of electron spin resonance scanning tunneling microscopy experiments has enabled the manipulation of individual adsorbate spin states physisorbed on ultrathin oxide layers supported on metal substrates. Electron resonance requires unpaired spin density on the adsorbate, which can be achieved, for instance, through charge transfer from the supporting substrate. This requires the correct energy-level alignment between the energy levels of the adsorbate and the Fermi energy of the substrate. Experiments on molecules and single atoms adsorbed on metal-insulator systems have revealed complex phenomena, including electronic bandgap narrowing, charge transfer, Fermi-level pinning, and the re-ordering of adsorbate orbitals after charge transfer. Despite these advances, a predictive first-principles approach based on accurate methods such as quasiparticle GW, capable of capturing these effects without the prohibitive cost of full adsorbate/oxide/metal simulations, remains an open challenge. In this work, we present a theoretical approach to determine the energy-level alignment of adsorbates on oxide/metal substrates. Our method transparently exposes all physical processes and strikes a balance between computational cost and accuracy. Ionization potentials and electron affinities of the isolated adsorbates are obtained using GW calculations, electronic bandgap polarization is quantified through the quasiparticle renormalization caused by the substrate, Fermi-level pinning is evaluated within the integer charge transfer model, and work function shifts arising from Pauli pushback or from the adsorbate-metal dipole are determined from the local variations of the electrostatic potential. This computationally efficient framework paves the way for highthroughput screening of molecular qubits and organic electronic interfaces.