Über die Elektronendynamik an helium- und graphenbedeckten Metalloberflächen

Die vorliegende Arbeit stellt eine experimentelle Studie zur Dynamik von Elektronen in Bildpotentialzuständen und in Grenzflächenzuständen an adsorbatbedeckten Metalloberflächen mittels zeitaufgelöster Zweiphotonen-Photoemissions-Spektroskopie(2PPE) dar. Dabei wurden zwei komplementäre Modellsysteme...

Full description

Saved in:
Bibliographic Details
Main Author: Armbrust, Nico
Contributors: Höfer, Ulrich (Prof. Dr.) (Thesis advisor)
Format: Doctoral Thesis
Published: Philipps-Universität Marburg 2012
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!

The present thesis represents an experimental study of the dynamics of electrons in image-potential states and in interfacial states at adsorbate covered metal surfaces by means of time-resolved two-photon photoemission (2PPE) spectroscopy. Two complementary model systems have been explored: On the one hand, a thin film of helium on a Cu(111) surface. Helium represents an ideal model for an insulating homogenous dielectric continuum because of its very low polarizability and its strong negative electron affinity. On the other hand, a graphene monolayer on top of a Ru(0001) substrate. Graphene exhibits a very high polarizability along the surface plane and gives rise to strongly bound series of image-potential states on its own. Moreover, the influence of a periodic corrugation of the graphene layer on the electronic structure at the combined system can be studied. The experimental investigation of the dynamics of electrons in image-potential states at the helium covered Cu(111) surface masters the challenges of combining this low-temperature experiment under ultrahigh vacuum conditions and the laser spectroscopy. It was possible to prepare a well-defined film of helium with a coverage of one monolayer on top of the Cu(111) substrate. It is found that thereby the binding energies of the unoccupied image-potential states n = 1 and n = 2 are strongly reduced compared to the clean Cu(111) surface. That of the (n = 1)-state is reduced by 50% and in the case of the (n = 2)-state it decreases by 30%. The reason for that is the strong decoupling of the image-potential states from the metal surface. It can be understood by the high tunneling barrier for the whole series of image-potential states which is presented by the helium adlayer as a result of the negative electron affinity in combination with the low polarizability of the helium. The increased distance of the states to the metal surface leads to the remarkable decrease of the binding energies. Thus, the (n = 1)-state also becomes an image-potential resonance as the higher states n = 2, 3, ... . As expected, the influence of the helium film results only in a relatively low energetic upshift of the partially occupied Shockley-surface state of the Cu(111) surface relative to the Fermi level. Furthermore, the lifetime of the first image potential state n = 1 shows a enormous increase by one order of magnitude. The lifetime of the (n = 2)-state is enhanced by a factor of 2 1/2. The explanation for this is that decoupling of the image-potential states is followed by a lesser interaction with bulk electrons. This reduces the inelastic decay channel. But primarily, the lifetime of the image-potential resonances is governed by the suppression of the decay by elastic electron transfer into the metal bulk. These results will be compared to calculations using one-dimensional model potentials which describe the film of helium by a tunneling barrier or a dielectric continuum. It will be shown that this can reproduce the trend of the binding energies. But they can not explain the change of the experimental lifetimes satisfactorily. Here, a more detailed model potential could lead quantitatively to a much better description. At the Ru(0001) surface, graphene forms a moiré superlattice which shows a remarkable periodic height modulation of 1.5 Å. This gives rise to a formation of surface areas with “hills” and “valleys” with different bonding lengths between the carbon atoms and the ruthenium substrate. At this surface, the first two image-potential states n = 1 and n = 2 can be observed. Compared to the clean Ru(0001) surface they show a lower binding energy and a sightly increased lifetime. This is an indication for a light decoupling effect by the graphene plane. By reason of the effective masses of these states, which are similar to that of a free electron, they are attributed to an almost freely moving series of image-potential states in the rather connected valley areas. Additionally, another image-potential state n = 10 can be observed which does not fit into the Rydberg-like series of the two other states. It exhibits a two times higher binding energy compared to the (n = 1)-state and a slightly shorter lifetime. It can be attributed to the hill areas. Due to the larger distance between the graphene and the metal substrate in these areas, the (n = 10)-state exhibits a considerable part of its probability density below the graphene sheet. The vicinity of the metal provokes a higher binding energy and a relatively short lifetime which is comparable to the clean Ru(0001) surface. Its much flatter dispersion is indicative for a stronger lateral location in the hill areas. Furthermore, two unoccupied states S0 and S can be observed 0.91 and 2.58 eV above the Fermi level. Their dispersion is close to that of a freely moving electron. Both originate from an unoccupied Shockley-type surface resonance. The low distance between the graphene and the substrate in the valleys causes a strong energetic upshift of this state. Therefore, it appears within the projected bandgap of the Ru(0001) surface forming an interfacial state S due to a possible hybridization with the first image-potential state. The trend of the energetic upshift is in accordance with that of other polycyclic organic molecules. For modeling the system, calculations have been carried out using a specially developed one-dimensional model potential. It basically describes the hills and the valleys on the basis of the different distances of the graphene plane. Qualitatively, the results explain the experimental findings very well. Moreover, the absolute values of the binding energies are reproduced satisfactorily. Taking many-body effects into account could help to reproduce the experimental lifetimes, as well