Sticking probability using different XC functionals in N2 /W(110)

G.A. BOCAN; M. ALDUCIN; R.DÍEZ MUIÑO; H.F. BUSNENGO; A. SALIN

San Sebastián, España

Workshop; Workshop on Elementary reactive processes at surfaces; 2007

Donostia International Physics Center

<!-- @page { margin: 0.79in } P { margin-bottom: 0.08in } --> Resumen The dynamics of simple molecules impinging on metallic surfaces is relevant for many technological applications, making its accurate theoretical description an important topic for research. The six-dimensional potential energy surface (PES) provides the basis for reasonably accurate results within the adiabatic frozen surface model and, in a recent article, such a PES has been obtained for the N2 / W(110) system using density functional theory and the corrugation reduction procedure [1]. Classical dynamics under the mentioned potential energy surface showed two distinct channels contributing to the dissociation process for normal incidence, namely, a direct one and an indirect one. The theoretical results for the sticking coefficient at low coverage compared well with experimental measurements at 800K, exception made of the range of low incident energies, where the indirect channel prevails. Also, a good fit to experiments was found for 60o incidence. Despite abundant literature suggesting dissipative effects as the main source of error in this kind of calculations we show here that the dissociation dynamics of N2 on W(110) is very sensitive to the functional chosen to model the exchange-correlation potential within the PES calculation. In particular, we compare the dissociation dynamics under the PW91 PES and the RPBE PES where PW91 and RPBE are two popular XC-functionals. The weight of the indirect channel is shown to be dramatically diminished in the RPBE calculation but also, its results for 60o incidence compare quite poorly with experiments. In fact our analysis leads us to conclude that the differences encountered between PW91 and RPBE results are to be taken as a measure of the DFT calculation accuracy for this kind of problems. [1] M. Alducin, R. Díez Muiño, H. F. Busnengo, and A. Salin, Phys. Rev. Lett. 97, 056102 (2006).