Andreev and spin transport in carbon nanotube quantum dot hybrid devices

In this thesis, we investigate carbon nanotube (CNT) quantum dot (QD) hybrid devices - in which a single CNT is coupled either to superconducting (S) and normal metal (N) electrodes or to ferromagnetic (F) contacts - by means of low-temperature transport spectroscopy experiments.
As a major technical prerequisite, we systematically develop and optimize different fabrication approaches with the aim to improve (i) the device yield and quality, and (ii) the for experiments with S or F contacts especially relevant CNT-metal interface cleanliness. To this end, an essentially residue-free e-beam lithography (EBL) based on ZEP520A resist proves particularly useful.
Using such fabrication techniques, we first study CNT QDs coupled to one S and either one or two N contacts in a two- or three-terminal device geometry. Due to the improved interface quality, optimized Pb- or Nb-based S contacts enable us to reliably obtain large and “clean” superconducting transport gaps of ~ 1 meV in S-QD devices also for varying couplings to the S contact, which poses a major advantage for the future study of S-QD hybrid devices. Due to this enhanced spectroscopic resolution for subgap bias voltages, we are able to analyze novel subgap transport phenomena. Depending on the coupling of the QD to the S contact and the other relevant energy scales, we investigate N-QD-S devices in three very distinct regimes, where different transport mechanisms dominate. For devices with weak coupling to S, the conductance is dominated by quasiparticle transport only, which allows us to demonstrate the impact of the gap on the Coulomb blockade diamond structure. Additionally occurring transport resonances at subgap bias voltages are either due to the thermal excitation of quasiparticles in S for elevated temperatures, or due to the peculiar three-terminal QD geometry and the extra transport channel between the two N contacts. For a device with an intermediate coupling to S, we identify for the first time spectroscopically resonant (elastic) and inelastic Andreev tunneling (AT) on a QD. This fundamental sequential transport process of two electrons through the same QD resonance provides the so far missing analogies to the Andreev reflection in metallic N-S structures and to the multiple Andreev reflections found in S-QD-S devices, and accounts for a competing local transport channel in Cooper pair splitter (CPS) devices. In devices with a sufficiently strong coupling to S, we investigate transport through Andreev bound states (ABS). Our peculiar three-terminal QD geometry enables us to identify the finite coupling to the N contacts as main source for the broadening of the visible Andreev resonances (AR), and to ascribe “excited” AR to the detailed energy level spectrum of the QD. Peculiar sign changes in the conductance through Andreev resonances between two N contacts allow us to qualitatively probe the gate-evolution of the ABS' Bogoliubov-de-Gennes amplitudes, and to identify the competition between the direct transport of a single electron through the AR (“resonant ABS tunneling”) and the non-local creation of a Cooper pair in S (or a Cooper pair splitting process) as origin of the sign changes. Our experiments with a floating S contact constitute a novel experimental probe for the superconducting proximity effect in S-QD systems, and potentially probe the strength of the coupling between S and the QD. Since we obtain mostly a single QD in our Pb-based three-terminal devices, we also propose and already demonstrate some first steps to reproducibly implement a double QD (DQD) in large-gap CPS devices. Such devices still open a wide range of experimental prospects. In a much wider context, our study of the N-QD-S model system also contributes to the continuing progress to understand subgap transport in superconductor hybrid devices, relevant also in the current quest to reveal and manipulate solid-state versions of Majorana fermions with a potential use in topological quantum computation schemes. For future experiments in such superconductor hybrid devices, the investigated three- or multi-terminal device geometry might prove extremely valuable. This geometry allows to unambiguously determine all contact tunnel couplings, and could hence give further insight into proximity-induced gaps in low-dimensional material systems with the help of a weakly coupled “density of states probe”, while simultaneously studying subgap transport with another probe.
In a second part of this thesis, we study CNT QDs coupled to two ferromagnetic contacts in a spin-valve device geometry, also to analyze the suitability of ferromagnetic leads as detectors of electron spin entanglement in CPS devices. With an optimized fabrication and characterization scheme, we obtain more reproducible magnetoresistance (MR) signals in these devices. Nevertheless, the consistently observed MR modulation on a negative MR offset stands in contrast to previous findings and orthodox theories of spin transport through QD devices, and is most likely due to the CNT-metal interface properties. Due to this still incoherent picture of spin transport in CNT QD spin-valves and the still not reproducible enough device characteristics, further optimization and experiments are needed to be able to employ ferromagnetic contacts for entanglement detection purposes, or other complex applications. To overcome these challenges, we propose to integrate the ferromagnetic contacts in ultra-clean processing schemes, and demonstrate some crucial first steps to achieve this challenging goal. A combination of these schemes with a CNT spin-valve structure possibly allows to investigate spin-transport in electrically stable devices with tunable tunnel barriers, or to perform Hanle-type experiments on a QD.
Altogether, the investigated QD hybrid devices with superconducting, ferromagnetic and normal metal contacts still provide a major playground to explore fundamental physics, or to come yet another step closer to applications of these devices in quantum technology.