Reproducibility of potential energy surfaces of organic/metal interfaces on the example of PTCDA on Ag(111)

The correct geometry is arguably key to studying the properties of molecules adsorbed on metallic substrates. The energetics of molecular adsorption depend on the interplay of various mechanisms: Covalent bonds, charge transfer, Pauli repulsion and van-der-Waals (vdW) interactions shape the potential energy surface (PES). Describing PESs with density functional theory (DFT) requires carefully selecting the exchange correlation functional and vdW correction. To explore the robustness of the PES with respect to the choice of the method, we benchmark common local, semi-local and non-local functionals in combination with various vdW corrections. We investigate these methods using perylenetetracarboxylic dianhydride (PTCDA) on Ag(111). [1]

We use an in-house developed machine learning algorithm, which requires only 50 DFT calculations as input to generate a PES with DFT accuracy. This allows analyzing features of the PES, such as positions and energies of minima and saddle points. PESs of different methods are in qualitative agreement, but also display significant quantitative differences. Lateral positions of adsorption geometries agree well with experiment, but adsorption heights, energies, and barriers show larger discrepancies.

[1] Hörmann et al., J. Chem. Phys. 153, 104701 (2020)