Comparison of dispersion-corrected exchange-correlation functionals using atomic orbitals

Abstract

for the adsorption of a noble-gas atom/monolayer as well as a graphene on the Cu(111), Pt(111), Pd(111), and Ag(111) close-packed surfaces. Most of these systems are characterized by weak interactions between the surface and the atom/monolayer, where the van der Waals interaction dominates. Here we investigate the above mentioned dispersion-corrected exchange-correlation functionals using atomic orbitals instead of plane waves applied in earlier works. We find that PBEsol+VV10s is an optimal functional for such hybrid interfaces, being accurate for both the equilibrium distance and the adsorption energy and providing realistic potential-energy curves. Moreover, we also show that PBEsol+VV10s and SG4+VV10m are both very accurate for the investigation of surface formation energies of all transition metals under consideration.

Friends or strangers? Attempts at reactivating buyer–supplier relationships

“…There, vdW-DF-cx provides results for the optimal adsorption distance that are closer to experimental observations [234] than what is obtained in RPA [235,236]. Reference [224] provides an additional comparison of vdW-DF-cx and of rVV10corrected semilocal functionals, including PBEsol + rVV10 [72], with the corresponding RPA results for graphene adsorption on the (111) surfaces of Cu, Pt, Pd, and Ag [235,236]. The vdW-DF-cx descriptions of the adsorption height are generally found in agreement with the RPA results for these surfaces [224].…”

“…We compute and extract the surface-energy and work function values for vdW-DF-cx, PBE, and PBEsol in QUANTUM ESPRESSO. Some SIESTA- [270] and VASP-based [217,218] vdW-DF-cx results for metal-surface work functions and surface energies have already appeared [224][225][226]. However, here we provide a broader vdW-DF-cx (and PBE/PBEsol) survey.…”

“…The case of weak chemisorption on Ag(111) merits a separate discussion, since the RPA results [235,236,279] for the adsorption height do not line up with recent measurements [280]. Reference [224] reports an adsorption height for vdW-DF-cx of 3.26 Å; adapting our reference [11] study, we find here a vdW-DF-cx value of 3.24 Å for graphene on Ag (111). These vdW-DF-cx results are in fair agreement with the RPA value [235,236], 3.31 Å, but not with the value 2.5 Å reported in reference [280].…”

“…The relevance of the first discriminator follows from the Hohenberg-Kohn theorem [221]: the structure-optimized nuclei positions R i must uniquely reflect the ground-state density variation n(r). The relevance of the third follows since the work function reflects the variation (across the surface) in the full electrostatic potential Φ(r) that acts on the electrons; the work functions and work-function shifts (computed for specific surfaces) is an established indicator of quality of XC functionals, for example, as discussed and used in references [25,89,[222][223][224][225][226].…”

“…Reference [224] provides an additional comparison of vdW-DF-cx and of rVV10corrected semilocal functionals, including PBEsol + rVV10 [72], with the corresponding RPA results for graphene adsorption on the (111) surfaces of Cu, Pt, Pd, and Ag [235,236]. The vdW-DF-cx descriptions of the adsorption height are generally found in agreement with the RPA results for these surfaces [224]. The vdW-DF-cx results for the adsorption energy per carbon atom also follows the RPA trend, but the values for vdW-DF-cx (and for the set of investigated rVV10-corrected functionals) are found systematically smaller than the RPA results.…”