Search for a command to run...
The combination of a topological insulator (TI) and a superconductor (S), which together form a TI/S interface, is expected to influence the possible surface states in the TI. It is of special interest, if the theoretical prediction of zero energy Majorana states in this system is verifiable. This thesis presents the experimental realization of such an interface between the TI strained bulk HgTe and the S Nb and studies if the afore mentioned expectations are met. As these types of interfaces were produced for the first time the initial step was to develop a new lithographic process. Optimization of the S deposition technique as well as the application of cleaning processes allowed for reproducible fabrication of structures. In parallel the measurement setup was upgraded to be able to execute the sensitive measurements at low energy. Furthermore several filters have been implemented into the system to reduce high frequency noise and the magnetic field control unit was additionally replaced to achieve the needed resolution in the μT range. Two kinds of basic geometries have been studied: Josephson junctions (JJs) and superconducting quantum interference devices (SQUIDs). A JJ consists of two Nb contacts with a small separation on a HgTe layer. These S/TI/S junctions are one of the most basic structures possible and are studied via transport measurements. The transport through this geometry is strongly influenced by the behavior at the two S/TI interfaces. In voltage dependent differential resistance measurements it was possible to detect multiple Andreev reflections in the JJ, indicating that electrons and holes are able to traverse the HgTe gap between both interfaces multiple times while keeping phase coherence. Additionally using BTK theory it was possible to extract the interface transparency of several junctions. This allowed iterative optimization for the highest transparency via lithographic improvements at these interfaces. The increased transparency and thus the increased coupling of the Nb’s superconductivity to the HgTe results in a deeper penetration of the induced superconductivity into the HgTe. Due to this strong coupling it was possible to enter the regime, where a supercurrent is carried through the complete HgTe layer. For the first time the passing of an induced supercurrent through strained bulk HgTe was achieved and thus opened the area for detailed studies. The magnetic dependence of the supercurrent in the JJ was recorded, which is also known as a Fraunhofer pattern. The periodicity of this pattern in magnetic field compared to the JJ geometry allowed to conclude how the junction depends on the phase difference between both superconducting contacts. Theoretical calculations predicted a phase periodicity of 4p instead of 2p, if a TI is used as weak link material between the contacts, due to the presence of Majorana modes. It could clearly be shown that despite the usage of a TI the phase still was 2p periodic. By varying further influencing factors, like number of modes and phase coherence length in the junction, it might still be possible to reach the 4p regime with bound Majorana states in the future. A good candidate for further experiments was found in capped HgTe samples, but here the fabrication process still has to be developed to the same quality as for the uncapped HgTe samples. The second type of geometry studied in this thesis was a DC-SQUID, which consists of two parallel JJs and can also be described as an interference device between two JJs. The DC-SQUID devices were produced in two configurations: The symmetric SQUID, where both JJs were identical, and the asymmetric SQUID, where one JJ was not linear, but instead has a 90° bent. These configurations allow to test, if the predicted uniformity of the superconducting band gap for induced superconductivity in a TI is valid. While the phase of the symmetric SQUID is not influenced by the shape of the band gap, the asymmetric SQUID would be in phase with the symmetric SQUID in case of an uniform band gap and out of phase if p- or d-wave superconductivity is dominating the transport, due to the 90° junction. As both devices are measured one after another, the problem of drift in the coil used to create the magnetic field has to be overcome in order to decide if the oscillations of both types of SQUIDs are in phase. With an oscillation period of 0.5 mT and a drift rate in the range of 5.5 μT/h the measurements on both configurations have to be conducted in a few hours. Only then the total shift is small enough to compare them with each other. For this to be possible a novel measurement system based on a real time micro controller was programmed, which allows a much faster extraction of the critical current of a device. The measurement times were reduced from days to hours, circumventing the drift problems and enabling the wanted comparison. After the final system optimizations it has been shown that the comparison should now be possible. Initial measurements with the old system hinted that both types of SQUIDs are in phase and thus the expected uniform band gap is more likely. With all needed optimizations in place it is now up to the successors of this project to conclusively prove this last point. This thesis has proven that it is possible to induce superconductivity in strained bulk HgTe. It has thus realized the most basic sample geometry proposed by Fu and Kane in 2008 for the appearance of Majorana bound states. Based on this work it is now possible to further explore induced superconductivity in strained bulk HgTe to finally reach a regime, where the Majorana states are both stable and detectable.