Comparison of the Performance of van der Waals Dispersion Functionals in the Description of Water and Ethanol on Transition Metal Surfaces

Abstract

Pairwise van der Waals (vdW) corrections have been routinely added to density functional theory (DFT) adsorption studies of inorganic or organic molecules on solid surfaces, however, comparative studies of the available pairwise corrections, e.g., D2, D3, D3(BJ), TS, and TS+SCS, are quite scarce. We report DFT calculations within the PerdewBurke-Ernzerhof (PBE) functional to assess the performance of the mentioned pairwise vdW corrections for well-defined transition-metal (TM) systems, namely, the Cu, Pt, and Au bulks in the face-centered cubic structure, close-packed TM substrates (Cu(111), Pt(111), Au(111), Cu-9/Pt-9/Cu(111), Pt-9/Cu-9/Cu(111), Au-9/Pt-9/Au(111), Pt-9/Au-9/Au(111)), and the adsorption of water and ethanol on the selected substrates, which include strained Pt-monolayers, i.e., a good challenge for pairwise vdW corrections. In general, accounting for vdW interactions leads to smaller lattice constants, which is expected due to the attractive nature of the vdW corrections, and the D3, D3(BJ), and TS+SCS improves the DFT-PBE results, in contrast with D2 and TS. Compared with PBE results, the vdW corrections enhance the contraction of the topmost surface layers, which contributes to change the electronic structure, in particular, the d-band center shifts away from the Fermi energy (up to 0.3 eV) in most cases, while the work function changes by about 0.2 eV in the worst cases. As expected, the attractive nature of the vdW corrections helps to enhance adsorption energies by 3-4 times compared with DFT-PBE. However, the adsorption energy trends versus the d-band center are preserved for all vdW corrections, except for the DFT-D2 framework, which deviates substantially from the studied vdW corrections. Therefore, based on our results and analyses, we can conclude that the D3, D3(BJ), and TS+SCS corrections yield the best description for the selected systems.