Fabrication and investigation of ultrahigh vacuum compatible interfaces of topological insulators and superconductors

The engineering of the lateral interface between a three-dimensional topological insulator and a superconductor for studying proximity effects is approached in two ways. Firstly, the fabrication of patterned samples using an ultrahigh-vacuum mask aligner is demonstrated. The piezomotor-driven mask aligner employs a silicon nitride shadow mask with capacitive sensors to read out the mask-sample distance. The fabrication process of such masks for the mask aligner is established. Consecutive evaporations can be made through this mask, with a horizontal sample position offset between each evaporation. This produces a superconducting layer with predefined holes on top of a topological insulator and provides access to the lateral interface via scanning tunneling microscopy (STM). Proof of principle for shadow mask evaporation using the mask aligner with evaporation of Au on Si(111) and Pb on Si(111) show sharp edges with penumbra less than 100 nm and partly down to 10 nm at safe mask-sample distances below 1μm without contact between them. Secondly, the influence of Nb on (Bi${}_x$Sb${}_{1-x}$)${}_2$Te${}_3$ is probed via angle-resolved photoemission spectroscopy (ARPES) and STM. Preliminary ARPES measurements indicate a shift in (Bi${}_x$Sb${}_{1-x}$)${}_2$Te${}_3$ bandstructure of 50 meV - 200 meV toward lower energies after the deposition of 0.3 monolayers of Nb. Patterned Nb-(Bi${}_x$Sb${}_{1-x}$)${}_2$Te${}_3$ covered with a protective Se-capping was optimized for STM. A superconducting gap in the tunneling spectrum on the Nb disappears at the interface to the (Bi${}_x$Sb${}_{1-x}$)${}_2$Te${}_3$, occasionally exhibiting a zero-bias peak. The tungsten tip indented into the Nb creates a superconducting microtip, resulting in a Josephson junction as confirmed by a supercurrent peak at zero bias and signature peaks of Andreev reflections. Conclusive STM results are impending since the fabrication process is not optimized as yet. Further pursuits for sample optimization require continuous feedback with STM characterization and exploring alternatives such as different capping materials.