Superconducting Proximity Effect

A normal metal in close contact with a superconductor will acquire superconductivity within a characteristic lengthscale given by the phase coherence length in the normal conductor. This proximity effect has been extensively studied in macroscopic samples and patterned two-dimensional electron gas structures. Carbon nanotubes are particularly interesting for investigating this phenomenon due to their long phase coherence lengths. Moreover, novel phenomena were predicted for 1D quantum wires in contact with superconductors [28].

Kazumov et al. [11] observed Josephson supercurrents in individual nano-tubes and ropes of nanotubes suspended between two superconducting bi-layer electrodes (gold/rhenium and gold/tantalum) when the devices were cooled below the superconducting transition temperature of the contacts. The nanotubes were embedded in the molten gold top layer (which becomes superconducting due to the proximity effect) to ensure low nanotube-electrode junction resistance - a key to observing the proximity effect. For ropes of nanotubes, the maximum supercurrent agreed with theoretical predictions of nA/eRN where A is the superconducting energy gap and RN the normal-state resistance of the junctions. However, the measurement on an individual nanotube gave a supercurrent that is 40 times higher than expected - a surprising result that has yet to be explained. In contrast, Morpurgo et al. [29] did not observe supercurrents through nanotubes sandwiched between two niobium contacts (Tc ~ 9.2 K) despite the fact that their samples had lower normalstate resistances. At 4.2 K, a backgate was used to tune the niobium-nanotube interface transparency. When the transparency is high, the transport across the interface is dominated by Andreev reflection processes (in which an incident electron is converted into a Cooper pair leaving a reflected hole in the normal region) which results in a dip around zero bias in the differential resistance. When the transparency is tuned low, the normal tunneling process dominates, resulting in a peak in differential resistance. The absence of a supercurrent and subharmonic gap structure due to multiple Andreev reflections, and the emergence of a sharp zero-bias differential-resistance peak superimposed on the Andreev dip at below ~4K seem to point to possible electron-electron correlation effects in carbon nanotubes.

0 0

Post a comment