Monday, October 17, 2011: 4:25 PM
Conrad D (Hilton Minneapolis)
Experiments have previously found that noble gases (Kr, Xe) adsorb on low-coordination atop sites on several different metal surfaces, rather than on high-coordination hollow sites. This unexpected preference for low-coordination sites has been previously ascribed to reduced Pauli repulsion, arising mostly from reduced exchange energy at the atop site. This explanation was based on density functional theory calculations within the local density approximation (LDA). In contrast, our calculations using he non-local van der Waals (vdW-DF) density functional show that site preference is due to a delicate balance between the electrostatics, which favor the hollow site, and kinetic energy which favors the atop site; exchange-correlation energies and, surprisingly, van der Waals interactions play a negligible role in site preference, although van der Waals interactions turn out to be critical in determining the magnitude of the adsorption energy and the correct equilibrium distance. Moreover, we find that the hollow site is a saddle point of index 1 or 2 (stationary points with 1 or 2 imaginary frequencies, respectively) on the 2-dimensional potential energy surface, while the atop site is the only true minimum. Therefore, the reason that hollow site occupation is not observed in experiments is that it is a transition state (see Figure 1) and so occupation of that site has a very short life-time. We show that this phenomenon holds for noble gas adsorption on close-packed Pt(111), Pd(111), Cu(111) surfaces, as well as the non-close-packed Cu(110) surface. Our results show that the inclusion of non-local vdW interactions is crucial for obtaining results in quantitative agreement with experiments for adsorption energies, equilibrium distances, and vibrational energies.
See more of this Session: Combining Weak and Strong Forces: Reactive Forcefields and Van Der Waals Enabled-DFT
See more of this Group/Topical: Computational Molecular Science and Engineering Forum
See more of this Group/Topical: Computational Molecular Science and Engineering Forum