CO adsorption on the Pt(111) surface: a comparison of a gradient corrected functional and a hybrid functional

Abstract

The adsorption of CO on the Pt(111) surface in a ( √ 3× √ 3) pattern has been studied with the gradient corrected functional of Perdew and Wang and the B3LYP hybrid functional. A slab which is periodic in two dimensions is used to model the system. The Perdew-Wang functional incorrectly gives the fcc site as the most favorable adsorption site, in accord with a set of previous studies. The B3LYP functional gives the top site as the preferred site. This confirms results from cluster studies where it was suggested that the different splitting, dependent on the functional, between highest occupied and lowest unoccupied molecular orbital, could be the reason for this change of the adsorption site. This is supported by an analysis based on the projected density of states and the Mulliken population.

CO oxidation on stepped-Pt(111) under electrochemical conditions: insights from theory and experiment

“…Indeed, CO adsorption on Pt surfaces is indeed a non-trivial system to be calculated with DFT in general as the approach usually tends to overestimate the 3-fold hollow sites over atop sites, which in low temperature experiments were found to be preferred. As our aim is to investigate the co-adsorption of CO and OH, here we point the reader to the extensive discussions about the so-called "CO/Pt(111) puzzle" [41][42][43][44][45][46][47][48][49][50][51][52][53].…”

Calculation of the work function with a local basis set

“…The Gaussian basis sets are: a [3s2p] basis set or a [4s3p] basis set for Li [10], [6s5p2d] basis sets for Cu [11] and Ni [6], a [4s3p2d] basis set together with a 19-valence electron pseudopotential for Ag [12], a [4s4p2d] basis set for Pt together with a 18-valence electron pseudopotential [13], and [5s4p1d] basis sets for Cl [11], and K [14]. k-point nets of the size 16 × 16 were used, and the smearing temperature was in the range between 0.001 E h and a maximum value of 0.01 E h , as described earlier [10,11,12,13,15]. In the case of nickel, spin-polarized calculations were performed, and the smearing temperature must be chosen low (0.001 E h ) because a too high temperature would artificially reduce the magnetic moment.…”

CO adsorption on metal surfaces: A hybrid functional study with plane-wave basis set

“…Such an increase of the surface energies is pointing towards an improved description using the hybrid functionals, but they are still underestimated compared to experiments: the experimental surface energies are ∼ 0.65 eV for Cu(111), ∼ 1.08 eV for Rh(111), ∼ 1.08 for Pt. 53 The B3LYP results 28,29 do not follow a consistent trend, with an increased surface energy for Cu and a decreased one for Pt. We have also calculated the PBE and HSE03 work function for the three (111) metal surfaces (not shown in Tab.…”

“…The underbinding has been attributed to the admixture of Fock exchange, and to the fact that Hartree-Fock usually underbinds, especially for metals. 28,43 We believe that this is not generally true for hybrid functionals. In fact, as shown in Ref.…”

“…As expected, using the standard LDA and GGA functionals the wrong site is preferred (shown in boldface), namely fcc, hcp and fcc for Cu, Rh and Pt in agreement with previous calculations. 28,29,57 The order of the sites with respect to the energy (starting from the most stable one) is fcc, hcp, top for Cu; hcp, fcc, top for Rh; fcc, hcp and top for Pt. These site orders generally agree with those reported in Refs.…”