Quantification of Atom Probe Tomography Data
Abstract
Among the characterization techniques which allow to "see" single atoms, atom probe tomography (APT) is unique in the way that individual atoms, instead of a collection of atoms, can be imaged while at the same time, the collected data can provide three dimensional information of atomic structures via reconstruction. In APT, individual or molecular species are field-evaporated from a sharp needle shaped specimen under an intense electric field and fly towards a 2D detector where their impact is recorded. The collective impacts form a "field desorption pattern" revealing the crystallographic structure of the specimen. While the patterns reveal the crystallography and atomic arrangement, they are not free of artefacts, and it is nearly impossible to quantify the resulting uncertainty in reconstruction due to the fully destructive nature of the technique. The major accomplishment in this project was the successful development of a predictive, hi-fidelity fully physical forward modeling code, "TAPSim-MD", for field evaporation. With this, we have explained artifacts in field desorption patterns such as high-density zone lines, artifacts in reconstruction from unexpected evaporation sequences, and the resolution limit of APT when it comes to placing atoms in the reconstruction, which can easily be approximately 2 nm, way larger than the atomic size. Along with TAPSim-MD, we have also developed a novel method to quantify bond energies from DFT calculations, which not only allows to understand surface binding and evaporation sequences, but as a side project allows thermodynamics of alloys and intermetallic on a simple intuitive, yet physically fully quantitative level. We finally have shown that our simulation data are suitable to train deep neural network-based reconstruction models that, when applied to experimental data, can indeed get APT down to atomic resolution and eliminate the worst artifacts.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 08, 2023
- Accession Number
- AD1231159
Entities
People
- Wolfgang Windl
Organizations
- Ohio State University