Frustrated H-Induced Instability of Mo(110)

Abstract

Using helium atom scattering Hulpke and Lüdecke recently observed a giant phonon anomaly for the hydrogen covered W (110) and Mo (110) surfaces. An explanation which is able to account for this and other experiments is still lacking. Below we present density-functional theory calculations of the atomic and electronic structure of the clean and hydrogencovered Mo (110) [3]) are only available for ΓH and ΓS. Thus, we focus our attention particularily on these two directions. For both surfaces two different softenings in the surface phonon branches were observed at the critical wave vectors: a smaller dip and a very deep and sharp indentation fromhω ≈ 15 meV tohω ≈ 2 meV. Only the phonon dispersion curves of quasi-one-dimensional conductors, like KCP [K 2 Pt(CN) 4 Br 3 ] [5] are characterized by similarly deep anomalies strongly localized in reciprocal space. A pronounced but less sharp damping of longitudinal surface phonons was observed for the (100) surfaces of W [6,7] and Mo [8] and was explained in terms of the nested structure of the Fermi surface [9,10]. In the case of W (100) this Fermi surface nesting is apparently so strong that it induces a rebonding at the surface and a structural rearrangement, i.e. a c(2 × 2) surface reconstruction at temperatures below T c ≈ 250 K (for a review see [11]). For the clean (110) surfaces of W and Mo no indication of an anomaly is found, but this only appears when the surfaces are covered with a full layer of hydrogen. Nevertheless, a close link between the surface phonon anomalies and H vibrations seems to be ruled out, since the HAS spectra remain practically unchanged when deuterium is adsorbed instead of hydrogen [2,3].Angular resolved photoemission (ARP) studies have been performed by Kevan and coworkers [12][13][14] for clean and hydrogen covered W (110) and Mo (110). These studies give no evidence of the existence of parallel segments of the Fermi-surface contours separated by wave vectors comparable with the HAS determined critical wave vectors and thus there appeared to be compelling reason for abandoning the nesting mechanism as a possible origin of the anomalies seen in HAS.The situation became even more puzzling when very careful high resolution electron energy loss spectroscopy (HREELS) studies [15] of hydrogen covered W (110) observed the Rayleigh wave phonon branch with the small dip at Q exp c1 exactly as the HAS results. However, the giant anomaly was not detected.At present there exists no explanation which is able to account for the observed two phonon anomalies and is compatible with the different experimental data (HAS, HREELS, and ARP). We therefore performed ab-initio calculations of the surface atomic geometries and the electronic properties of the clean and H-covered Mo (110) surface. The results for the adsorbate covered surface provide clear evidence for a Fermi surface nesting instability at wave vectors which are in remarkable agreement with those at which the phonon anomalies occur for this system [Q . We choose the x-axis along [001] and t...