Spectroscopic investigations of proximity-systems: setup of a 300-mK scanning tunneling microscope with high energy resolution and investigations of the density of states on the system aluminum/gold

The work describes the setup of an ultra low temperature Scanning Tunneling Microscope (STM), which works in a conventional ³He-cryostat with a base temperature of 240 mK and a vector magnet with a magnetic field up to 1 T. The STM is rigid and compact and therefore it is insensitive to vibrations and can be used without elaborate damping. The STM is well shielded against radio-frequency radiation and all electrical lines are filtered at low temperatures. The measurement of the differential conductance dI/dU of superconducting aluminum yields a temperature of the electronic system of 270 mK and a energy resolution of 100 µeV. The STM is nonmagnetic, what has been demonstrated on the successful observation of superconductivity of aluminum, which has a critical field of only 10 mT. Furthermore the characteristic hysteresis in the phase transition between the normal- and the superconducting state could be observed on a aluminum tip. The atomic resolution of graphite at 240 mK in a magnetic field of 1 T shows the capability to work in high fields as well.The investigation of the Proximity-Effect has been performed with the developed STM using tunneling spectroscopy and recording the differential conductance dI/dU on the System aluminum/gold. The temperature dependence as well as the dependence on the magnetic field was investigated. The thickness of the superconductor aluminum was 320 nm and the thickness of the normal metal gold has been varying from 20 nm up to 190 nm. The induced minigap decreases slower with increasing thickness of the normal metal, than predicted by the theory. Based on experimental values of the mean free path, the differential conductance is calculated in the frame work of the quasiclassical theory. In contrast to the experimental curves the theoretical curves show a significantly smaller minigap. The Variation of parameters like the interface resistance and the coupling parameter between superconductor and normal metal do not result in a better description of the experimental data. Even the assumption of a reduced normal metal thickness yields no agreement regarding the temperature dependence. Only the assumption of a ten times larger thickness of the superconductor results in a satisfied description of the dI/dU-curves. Also the temperature dependence can be described within this assumption.The Experiments and the required increase of the superconducting layer indicate that the superconductivity in aluminum is less suppressed by the normal metal gold, than predicted by the theory in the case of a thin film