Spin and heat transport in nanoscale superconductor hybrid structures

In the main part of this work the first experimental observations of thermoelectric effects in superconductor/ferromagnet tunnel junctions were treated. For this purpose, samples were fabricated which had a six-probe tunnel junction, so that thermoelectric current could be measured simultaneously to the generation of a temperature gradient across the junction. For the measurements a second harmonic detection technique was established and during the experiments it was checked that the results derived by this technique were real and not induced by spurious contributions. The successful detection of thermoelectric currents showed that the breaking of the electron-hole symmetry of the conductance in spin-polarized superconductor tunnel junctions leads to large thermoelectric effects in the presence of a spin-splitting field and confirmed the theoretical predictions. By extracting the necessary junction parameters and the heater calibration from local conductance measurements, we were able to model the theoretically expected current quantitatively in applied magnetic field and observed excellent agreement between experiment and theory. The inferred Seebeck coefficients reached magnitudes around -100 μV/K, much larger than it is typically observed in metallic structures. The absence of the thermoelectric effect in a nonmagnetic sample of the same design showed that the spin polarization of the junction is crucial for the effect and that spin and heat currents are coupled in high-field superconductor hybrid structures. By measurements of the dependence of the thermal excitation on the thermoelectric current we observed that the thermoelectric signal has a slight nonlinearity in small fields, but is almost completely linear in the gapless region of the superconducting state. Hence, this regime seems promising for high-sensitive electron thermometry in improved devices. Subsequently, the experiments were extended to ferromagnetic tunnel junctions with superconductors coupled in proximity contact to a ferromagnetic insulator. The thermoelectric effect itself was measurable in the structures and by including the exchange field to the model we observed again excellent agreement between the measured data and the theoretically expected current. The author could show that by the use of an ideal ferromagnetic junction the FIS junction would outperform the NIS junction for smaller voltages due to the additional contribution of the linear thermoelectric effect. Finally, the first measurements of an experiment which has the intention to study the Hanle precession of the quasiparticle spin projection in spin-split superconductors were presented.