Kurt Gloos

Electron and hole transmission through superconductor — Normal metal interfaces

We have investigated the transmission of electrons and holes through interfaces between superconducting aluminum (Tc = 1.2K) and various normal non-magnetic metals (copper, gold, palladium, platinum, and silver) using Andreev-reflection spectroscopy at T = 0.1K. We analysed the point contacts with the modified BTK theory that includes Dynes’ lifetime as a fitting parameter Γ in addition to superconducting energy gap 2Δ and normal reflection described by Z. For contact areas from 1 nm2 to 10000nm2 the BTK Z parameter was 0.5, corresponding to transmission coefficients of about 80%, independent of the normal metal. The very small variation of Z indicates that the interfaces have a negligible dielectric tunneling barrier. Fermi surface mismatch does not account for the observed transmission coefficient.

Andreev-reflection spectroscopy of ferromagnets: The impact of Fermi surface mismatch

We have investigated point contacts between superconductors (Nb, AuIn2) and normal metals (ferromagnetic Co, nonmagnetic Cu). The observed Andreev-reflection spectra are analyzed using a modified BTK theory including spin polarization effects. This results in a polarization of Co that agrees with observations by others, but lifetime effects describe the spectra equally well. On the other hand, the spectra with nonmagnetic Cu can be described well by a spin-polarization model. The discrepancy between the polarization and lifetime interpretations poses a dilemma which can be resolved by considering normal reflection at those interfaces owing to Fermi surface mismatch. Our data suggest that Andreev reflection at Nb-Co contacts gives the correct magnetic polarization of Co only when lifetime effects and intrinsic normal reflection are included.

An alternative view at the zero-bias anomaly of metallic point contacts

We have studied the zero-bias anomaly of point contacts as function of contact size for a wide range of materials from simple normal metals, superconductors, and magnets to highly correlated heavy-fermion compounds. The size δR of the zero-bias anomaly varied in a universal manner proportional to the square of the contact resistance R between 1Ω and 10kΩ, that is for contact radii from about 15nm down to 0.2nm. Magnetic impurities, two-level systems, or a tunneling barrier are unlikely to be the main source of this anomaly. We suggest instead Kondo-type scattering of the conduction electrons due to a spontaneous electron spin polarization at the point contact.

Anisotropic multiband superconductivity in Zn observed by simultaneous Andreev reflection and Yanson point-contact spectroscopy

Andreev reflection (AR) and Yanson point-contact spectroscopy (PCS) have been applied simul-taneously to study the superconducting (SC) gap and the electron-phonon interaction (EPI) in a Zn single crystal. The correlation between SC gap value and EPI spectrum allowed us to establish the anisotropy of the SC gap. Evidence for multiband superconductivity in Zn is present with two gaps related as 1:1.3. We also found that the AR features are more robust against the point-contact quality than the EPI ones, possibly, due to the large coherence length in Zn compared to the typical PC size. Even for the PCs close to the ballistic regime with intense EPI spectra, the transmission coefficient evaluated from the AR spectra is near the one predicted theoretically for the diffusive regime of a current flow in the PC. Understanding those phenomena would provide a more reliable basis to apply both types of spectroscopies to study more complex SCs.

Anatomy of point-contact Andreev reflection spectroscopy from the experimental point of view

We review applications of point-contact Andreev-reflection spectroscopy to study elemental superconductors, where theoretical conditions for the smallness of the point-contact size with respect to the characteristic lengths in the superconductor can be satisfied. We discuss existing theoretical models and identify new issues that have to be solved, especially when applying this method to investigate more complex superconductors. We will also demonstrate that some aspects of point-contact Andreev-reflection spectroscopy still need to be addressed even when investigating ordinary metals.

Phonon-drag induced suppression of the Andreev hole current in superconducting niobium contacts

We have investigated how the Andreev-reflection hole current at ballistic point contacts responds to a large bias voltage. Its strong suppression could be explained by the drag excerted by the non-equilibrium phonon wind generated by high-energy electrons flowing through the contact. The hole - phonon interaction leads to scattering lengths of the low-energetic holes down to 100\,nm, thereby destroying the coherent retracing of the electron path by the Andreev-reflected holes.

Break-junction experiments on the zero-bias anomaly of non-magnetic and ferromagnetically ordered metals

We have investigated break junctions of normal non-magnetic metals as well as ferromagnets at low temperatures. The point contacts with radii 0.15—15 nm showed zero-bias anomalies which can be attributed to Kondo scattering at a single Kondo impurity at the contact or to the switching of a single conducting channel. The Kondo temperatures derived from the width of the anomalies varied between 10 and 1000 K. These results agree well with literature data on atomic-size contacts of the ferromagnets as well as with spear-anvil type contacts on a wide variety of metals.

Spin polarization versus lifetime effects at point contacts between superconducting niobium and normal metals

Point-contact Andreev reflection spectroscopy is used to measure the spin polarization of metals but analysis of the spectra has encountered a number of serious challenges, one of which is the difficulty to distinguish the effects of spin polarization from those of the finite lifetime of Cooper pairs. We have recently confirmed the polarization-lifetime ambiguity for Nb-Co and Nb-Cu contacts and suggested to use Fermi surface mismatch, the normal reflection due to the difference of Fermi wave vectors of the two electrodes, to solve this dilemma. Here we present further experiments on contacts between superconducting Nb and the ferromagnets Fe and Ni as well as the noble metals Ag and Pt that support our previous results. Our data indicate that the Nb - normal metal interfaces have a transparency of up to about 80% and a small, if not negligible, spin polarization.

Normal reflection at superconductor - normal metal interfaces due to Fermi surface mismatch

Electrons can be reflected at an interface between two metals because of a dielectric barrier or different properties of the Fermi surface. Andreev reflection allows to directly measure normal reflection when one of the metals is a superconductor. We have investigated normal reflection at interfaces between non-magnetic normal metals and superconducting Nb (Tc = 9.2 K) and Al (Tc = 1.2 K). The distribution of the values of the relative strength of the interface barrier, Z, for a number of contacts of a specific metal combination shows a well-defined peak which can be attributed to Fermi surface mismatch. Our reflection coefficients are generally smaller than those predicted theoretically or those derived from proximity-effect studies of normal-superconductor bi-layers.

The zero-bias anomaly of point contacts with ferromagnetic Ni and with the heavy-fermion antiferromagnet CeAl2

We have investigated spear-anvil type point-contacts between ferromagnetic nickel as well as the heavy-fermion antiferromagnet CeAl2 and various simple metals (Cu, Ta, Nb). Contacts with small resistance usually showed electron-phonon scattering, Andreev reflection in case of superconducting counter-electrodes, as well as anomalies due to magnetic ordering. With increasing contact resistance (decreasing contact size) a zero-bias anomaly appeared in both Ni and CeAl2 contacts. It is conventionally attributed to resonant scattering at two-level systems or at magnetic impurities (Kondo effect). At contacts of ~ 1 nm diameter it suppressed completely all other spectral features. We discuss whether those mechanisms are relevant here and what alternatives there might be.