Modeling of Ballistic Electron Emission Microscopy on metal thin films

TL;DR

This paper develops an extended theoretical model for Ballistic Electron Emission Microscopy (BEEM) to better understand electron transport in thin metal films and heterostructures, crucial for spintronics applications.

Contribution

It introduces an iterative Green function calculation method for finite systems and a simplified approach that approximates non-equilibrium effects in BEEM modeling.

Findings

Extended model to thin films and heterostructures

Simplified method approximates non-equilibrium results

Enables detailed analysis of layer interface effects

Abstract

After the discovery of GMR by Fert and Gr\"unberg, electronics had a breakthrough with the birth of a new branch called spintronics. This discipline, while still young, exploits the spin of electrons. Most quantum devices exploiting this property of electrons consists of alternating magnetic and nonmagnetic thin layers on a semiconductor substrate. One of the best tools used for characterizing these structures is the so-called Ballistic Electron Emission Microscope (BEEM). Originally, it was dedicated to the imaging of buried objects and to the study of the potential barrier formed at the interface of a metal and a semiconductor. With the development of spintronics, the BEEM became an essential spectroscopy technique but still fundamentally misunderstood. The first realistic model, based on the non-equilibrium Keldysh formalism, was proposed to describe the transport of electrons during BEEM experiments. However, despite its success, its use was limited to the study of semi-infinite structures through a calculation method called decimation of Green functions. In this context, we have extended this model to the case of thin films and hetero-structures like spinvalves: starting from the same postulate that electrons follow the band structure of materials in which they propagate, we have established an iterative formula allowing calculation of the Green functions of the finite system by tight-binding method and encoded it in a F90 code. In parallel, we have developed a simpler method which allows to avoid passing through the non-equilibrium Keldysh formalism. Despite its simplicity, we have shown that this intuitive approach gives some physical interpretation qualitatively similar to the non-equilibrium approach. However, for a more detailed study, the use of "non-equilibrium approach" is inevitable, especially for the detection of thickness effects linked to layer interfaces.