Thermoelectric Modeling of the Non-Ohmic Differential Conductance in a Tunnel Junction containing a Pinhole

TL;DR

This paper demonstrates that fitting differential conductance to tunneling models cannot reliably detect pinhole defects in tunnel junctions, as significant current may bypass tunneling paths.

Contribution

It introduces a model for electrical and thermal properties of tunnel junctions with pinholes, revealing limitations of traditional fitting methods.

Findings

Fitting conductance does not confirm junction quality when pinholes are present.

Up to 88% of current can bypass tunneling through pinholes.

Traditional models may falsely indicate good junctions despite defects.

Abstract

To test the quality of a tunnel junction, one sometimes fits the bias-dependent differential conductance to a theoretical model, such as Simmons's formula. Recent experimental work by {\AA}kerman and collaborators, however, has demonstrated that a good fit does not necessarily imply a good junction. Modeling the electrical and thermal properties of a tunnel junction containing a pinhole, we extract an effective barrier height and effective barrier width even when as much as 88% of the current flows through the pinhole short rather than tunneling. A good fit of differential conductance to a tunneling form therefore cannot rule out pinhole defects in normal-metal or magnetic tunnel junctions.