Regensburg 2016 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
O: Fachverband Oberflächenphysik
O 71: Scanning Probe Techniques and New Experimental Methods
O 71.16: Poster
Mittwoch, 9. März 2016, 18:15–20:30, Poster A
Reliability of LEED analyses for complex systems — •Lutz Hammer, Pascal Ferstl, and M.Alexander Schneider — Lehrstuhl für Festkörperphysik, Universität Erlangen-Nürnberg
The correspondence of experimental and calculated best-fit intensity spectra of a LEED structure analysis is quantitatively expressed by means of reliability factors. Most frequently the Pendry R-factor RP is applied nowadays. Best-fit values in the range RP = 0.2 - 0.3 or even somewhat larger are generally accepted as a proof for the correctness of the underlying structural model. In the present contribution it is shown that this only holds for structurally quite simple systems. For more complex structures, in contrast, the distinction between similar models which only vary by the presence (or absence) of very few atoms within the (large) unit cell requires a significantly better level of spectral correspondence (RP ≪ 0.2).
In the case study presented here different structural variations of the surface reconstruction of Fe3O4(100) are investigated, whereby the key features of the reconstruction model (subsurface iron vacancies and interstitials) are systematically either neglected or varied in position. It is shown that those crystallographically wrong structures still produce R-factors in the range RP = 0.2 - 0.3, occasionally even lower. The reason for this structural insensitivity lies qualitatively in the relatively small scattering contribution of a particular atom to the total wave field produced by a large surface unit cell. Approaches to improve the quality of a LEED intensity analysis as well as consistency checks to detect incomplete structural models are presented and discussed.