A novel technique for tunnel junction diagnostics

We present a way to diagnose problems in the preparation of tunnel junctions by numerical deconvolution of the differential conductance data. The conductance of a junction can deviate from the ideal BCS conductance by a broadening in the gap region. The broadening can be due to several causes: The smearing can be caused by the presence of two distinct phases with two different energy gaps or by a continuous distribution of energy gaps, due to either anisotropy effects or proximity effect from a contaminated surface. These cases may not be distinguishable in a measurement at finite temperature. But after the thermal smearing is removed by numerical deconvolution, the details in the distribution of energy gaps can be restored. This has important applications for the study of junctions on transition metals and especially on high Tcternary superconductors like Chevrel phases and rhodium borides, where these methods may reveal features in the density of states, which are not apparent in the conductance data because of thermal smearing. The performance and potential of numerical deconvolution methods are discussed and demonstrated on tunneling data of a V3Ga-junction in a magnetic field.