X-ray Computed Tomography (XCT) is an often used non-destructive inspection tool. To identify small features and determine whether they are real or an image artefact, it is important to know the spatial resolution of the dataset. The typical rule-of-thumb for spatial resolution of a CT scan suggests a feature must be three voxels wide to be observed, but this is strongly dependent on scan quality. For a broad range of applications this estimate is sufficient but when working at the limits of the system capabilities numerical verification of the spatial resolution per scan would be preferred. Currently the most used verification method for spatial resolution is the JIMA chart but this is a 2D radiographic method that does not directly translate to the 3D reconstructed volume. This research identifies micro-manufacturing methods that can be used to make 3D spatial resolution verification objects, overcoming the limitations of macro manufacturing with small defined features but a large volume. Two different designs have been explored: intrusions and extrusions to represent both internal features such as pores and external features respectively. For simplicity, both are based on equidistant line pairs as per their 2D counterparts. For the intrusions PFIB was used to etch the features, while for the extrusions two photon polymerisation (2PP) and an electro chemical printing method were used. The research showed that while manufacturing volume and small features is still challenging, it is possible to make an object with walls in both a metal and a polymer with stable feature sizes down to 5 μ m . The 2PP method is particularly promising with aspect ratios down to 1:10, the relatively low cost and ease of access. The objects manufactured in this study can simply be replicated by other XCT users to verify the 3D spatial resolution in their scans, and can further be tailored for their sample.
Zwanenburg et al. (Sun,) studied this question.