Using alpha hulls to automatically and reproducibly detect edge clusters in atom probe tomography datasets

Abstract

Highlights  Novel method developed for automatically detecting edge clusters in atom probe datasets  Application to simulated data and its effect on measurement accuracy is assessed  Successfully applied to real dataincluding datasets with complex shapes

Observation of Mn-Ni-Si-rich features in thermally-aged model reactor pressure vessel steels

“…Clusters that intersected the edge of the dataset were identified using the methods outlined in Ref. [28] edge clusters were calculated as half for number density (Nd) calculations. Due to the small number of detected clusters in each state, edge clusters were included in composition and size calculations.…”

Atom probe characterisation of segregation driven Cu and Mn–Ni–Si co-precipitation in neutron irradiated T91 tempered-martensitic steel

“…Partial edge clusters (defined as a cluster from the original material whose true extent is not completely contained within the reconstructed APT volume) were removed from the cluster search to avoid underestimation of the sizes. The algorithm used to detect and remove the edge clusters was developed by Jenkins et al (2020) [58,69]. The number density of clusters, N !…”

Alpha Shape Analysis (ASA) Framework for Post- Clustering Property Determination in Atom Probe Tomographic Data

“…In this work, alpha shapes are identified and selected as an approach for determining the concave hull. Alpha shapes have already been used to identify a mesh for the overall atom probe specimen in order to improve reconstructions as in Felfer & Cairney (2016) or to identify clusters existing near the boundary of the data as in Jenkins et al (2020). Additionally, we propose the alpha shape analysis (ASA) framework which enables irregular meshing of complex point clouds through a stochastic cluster-specific empirical heuristic for selection of its singular input parameter, α , which describes the radius of a carving sphere.…”