Applying computational geometry techniques for advanced feature analysis in atom probe data

•Computational geometry is used to detect and analyse features within atom probe data.•Limitations of conventional feature detection are overcome by using atomic density gradients.•0D, 1D, 2D and 3D features can be analysed by using Voronoi tessellation for spatial binning.•New, robust analysis methods are demonstrated, including line and interfacial excess mapping.

In this paper we present new methods for feature analysis in atom probe tomography data that have useful applications in materials characterisation. The analysis works on the principle of Voronoi subvolumes and piecewise linear approximations, and feature delineation based on the distance to the centre of mass of a subvolume (DCOM). Based on the coordinate systems defined by these approximations, two examples are shown of the new types of analyses that can be performed. The first is the analysis of line-like-objects (i.e. dislocations) using both proxigrams and line-excess plots. The second is interfacial excess mapping of an InGaAs quantum dot.