Comparison of the periodic slab approach with the finite cluster ansatz for metal-organic interfaces at the example of PTCDA on Ag(110)

TL;DR

This study compares the effectiveness of periodic slab and cluster models in predicting metal-organic interface properties, finding that larger clusters with four silver layers closely match slab results, offering a viable alternative for complex surface systems.

Contribution

It demonstrates that sufficiently large cluster models can accurately reproduce slab results for metal-organic interfaces, providing a practical alternative for non-periodic surface studies.

Findings

Clusters with four silver layers reproduce slab adsorption energies within 10%.

Larger clusters accurately match the structure within 0.02 Å.

Cluster approach is computationally competitive for defect and non-periodic systems.

Abstract

We present a comparative study of metal-organic interface properties obtained from dispersion corrected density functional theory calculations based on two different approaches: the periodic slab supercell technique and cluster models with 18 to 290 Ag atoms. Fermi smearing and fixing of cluster borders are required to make the cluster calculation feasible and realistic. The considered adsorption structure and energy of a PTCDA molecule on the Ag(110) surface is not well reproduced with clusters containing only two metallic layers. However, clusters with four layers of silver atoms and sufficient lateral extension reproduce the adsorbate structure within 0.02 \AA\ and adsorption energies within 10\% of the slab result. A consideration of the computational effort shows that the cluster approach is a competitive alternative to methods using periodic boundary conditions and of particular interest for research at surface defects and other systems that do not show periodic symmetry.