Microstructural analysis of TRISO particles using multi-scale X-ray computed tomography

T. Lowe, R.S. Bradley, S. Yue, K. Barii, J. Gelb, N. Rohbeck, J. Turner, P.J. Withers - Manchester X-ray Imaging Facility, School of Materials, University of Manchester, UK ; School of Mechanical Engineering, University of Manchester, UK ; Zeiss Xradia Inc., Pleasanton, CA, USA ; The Research Complex at Harwell, Rutherford Appleton Laboratory, Didcot, Oxfordshire, UK

TRISO particles, a composite nuclear fuel built up by ceramic and graphitic layers, have outstanding high temperature resistance. TRISO fuel is the key technology for High Temperature Reactors (HTRs) and the Generation IV Very High Temperature Reactor (VHTR) variant.

TRISO offers unparalleled containment of fission products and is extremely robust during accident conditions. An understanding of the thermal performance and mechanical properties of TRISO fuel requires a detailed knowledge of pore sizes, their
distribution and interconnectivity. Here 50 nm, nano-, and 1 lm resolution, micro-computed tomography (CT), have been used to quantify non-destructively porosity of a surrogate TRISO particle
at the 0.3–10 lm and 3–100 lm scales respectively. This indicates that pore distributions can reliably be measured down to a size approximately 3 times the pixel size which is consistent with the segmentation process. Direct comparison with Scanning Electron Microscopy (SEM) sections indicates that
destructive sectioning can introduce significant levels of coarse damage, especially in the pyrolytic carbon layers. Further comparative work is required to identify means of minimizing such damage for SEM studies. Finally since it is non-destructive, multi-scale time-lapse X-ray CT opens the possibility
of intermittently tracking the degradation of TRISO structure under thermal cycles or radiation conditions in order to validate models of degradation such as kernel movement. X-ray CT in-situ experimentation of TRISO particles under load and temperature could also be used to understand the internal changes that occur in the particles under accident conditions.

How Amira-Avizo Software is used

The reconstructed data was then analysed using Avizo standard
7.0  to segmentand visualise the virtual slices and 3D volume renderings. For the Versa XRM datasets the porosity in the Buffer, IPyC and OPyC layers was segmented using the adaptive thresholding technique as described in Section 3.3. As the width of the pores is relatively narrow and pore morphology fairly consistent, a baseline image was created using a morphological closing filter (dilation followed by erosion of same kernel size). This process averages the grey scale values of neighbouring voxels so that the pores could be
adaptively extracted by subtracting the original image from the filtered
image resulting in a pore outline.