Photoemission of highly correlated systems

Construction of a new UHV chamber for high resolution photoemission spectroscopy. An UHV-chamber with a high resolution analyzer was constructed. The analyzer was tested with photoemission spectroscopy on argon gas. The maximum resolution reached was 10 meV with a pass energy of 1eV. An angular resolution of ∼ ± 0.65o was demonstrated by measuring the Cu(110) surface state at Y. Pt(110). Angle-resolved photoemission spectroscopy (ARPES) of Pt(110) reveals features which cannot be interpreted as a result of normal band dispersion. An emission at X is very likely due to a surface state. The abnormally high intensity and the similarity with ARPES line shapes of highly correlated systems suggests that here many-body effects are observed even on the clean metal. A possible explanation arises from a symmetry analysis, according to which the surface state is probably a derived one-dimensional Tamm state. Bromine and chlorine on Pt(110). With a small additional adsorption of bromine or chlorine both, the c(2x2)-Br/Pt(110) surface and the p(2x1)-C1/Pt(110) surface show a phase transition into a (3x1) phase, which is interpreted as a charge density wave (CDW). The formation of a CDW is favored by the presence of one dimensional electronic state and is associated with the opening of a band gap at the Fermi level. ARPES-spectra of the system c(2x2)-Br/Pt(110) show both, one dimensional behavior and a band umklapp as well as the lowering of the upper band edge below EF. ARPES-spectra: highly correlated electronic systems. ARPES-spectras of the quasi-1D system Br/Pt(110) show features, which significantly differ from those expected in a normal photoemission process. Narrow peaks with dramatic intensity enhancement upon approaching the Fermi energy (quasi-particle peaks), strong renormalization of the effective mass and the anomalous temperature dependence of some quasi-particle peaks signal pronounced correlation effects. Strong electron-phonon-interaction as a possible cause for the observed spectral functions can be dismissed: evaluation of the real part of the self energy in the system Br/Pt(110) would indicate an energy of more than 200 meV for the interacting phonons. However such phonons can be excluded in the system Br/Pt(110) because of the large mass of the constituent particles. The ARPES-spectras of the quasi-1D c(2x2)-Br/Pt(110) surface show a striking similarity in peakshapes and temperature dependence with the spectra of the high TC cuprate Bi2Sr2CaCu2O8. Within the framework of the Luttinger liquid model the quasi-particle peak in the spectra of the high TC cuprates is interpreted as a bound state of two interacting quasi-particles (spinon and holon). If the attracting interaction of spinon and holon is k-dependent , the line shape of the quasiparticle-peak varies with k. The disappearance of the quasiparticle peak at T + 62TC is caused in this model by the disappearance of the attractive interaction of spinon and holon due to a dimensional crossover from 2D to 1D. In the system c(2x2)-Br/Pt(110) such a dimensional crossover from 2D to 1D was demonstrated by LEED-measurements. This result supports the dimensional-crossover model for the high TC cuprates. Fermi surface engineering with NO. For c(2x2)-Br/Pt(110) the critical Fermi surface giving rise to the CDW is located at kF ∼ G/3, which leads to a CDW with a threefold periodicity. Adsorption of NO, too, yields a (3x1)-structure with similar electronic structure as the (3x1)-Br/Br/Pt(110) surface. Additional adsorption of NO induces a shift of the Fermi vector kF. Extrapolation of the band dispersion for an NO-coverage of 0.25ML results in a Fermi vector kF ∼ G/4, which should lead to a CDW with twofold periodicity. However, a (4x1)-structure was observed. This is caused by NO(CO) zigzag-rows, which due to direct adsorbate-adsorbate interaction cannot approach each other below a critical distance of 4 x 2.77 A. Measurements on platinum powder pellets. In contrast to bulk-platinum, platinum powder becomes superconducting at low temperatures. A possible explanation would be a surface-related superconductivity. This hypothesis can be examined by modification of the surface. Bromine induces different phases on the Pt(110) surface, so it might be interesting to measure the conductivity of pure and brominated platinum powder pellets. Pure platinum powder pellets show the expected linear decrease of the resistance as the temperature is lowered. On brominated powder pellets a significant reduction of the resistivity is observed in the first measurement after the bromination. Repetition of the measurement on the same sample always yielded a partial recovery towards higher resistivity. The increase of the conductivity after the bromination obviously demonstrates a strong influence of the surface on the total (bulk and surface) conductivity. The increase of the resistivity following the first measurement after the bromination could be caused by the condensation of water on the sample. Prior to definitive conclusions, side effects like sintering and water condensation have to be brought under control. (author)