Katie Jones



The pore-level characterisation of nuclear graphite is critical for predicting reactor safety and for assessing the viability of different grades of graphite material. The majority of studies often focus on the impact of one specific length scale (macroscopic/mesoscopic/ microscopic) but very few studies have attempted to provide void size information which spans multiple length scales. This study, therefore, aimed to advance the knowledge of the entire void range for Gilsocarbon graphite, including any changes to the void structure that occur as a consequence of irradiation damage and radiolytic oxidation, by incorporating a combination of experimental and modelling techniques. Gilsocarbon graphite, which is incorporated in the current UK Advanced gas cooled reactors, is particularly difficult to characterise at a pore level due to its highly complex pore matrix which comprises voidage over several orders of magnitude. Additionally, the radioactive nature of the samples limits the amount of material available for analysis. In order to successfully measure the small sample volumes provided, novel instrumentation and interpretation methods were developed. This included the construction of a micropycnometer, built to obtain values of the accessible and inaccessible pore volumes. In addition, graphite's low surface area demanded the use of krypton as an adsorbative, which required the acquisition of a high performance instrument as well as the development of an interpretive GCMC kernel for obtaining pore size information. These data were used to correct the pore size information obtained at high pressure during mercury porosimetry, as the porosimetry data was suspected to contain inaccuracies due to damage or deformation of the graphite's microstructure caused by the analysis. The bespoke software package PoreXpert, designed at the University of Plymouth, was used to inverse model the experimentally measured percolation characteristics and total accessible porosity to generate simulated void network structures. The improved, quasi-Bayesian, modelling of the combined percolation curves identified differences in the pore size distributions for Gilsocarbon samples during various stages of ageing. The findings from the bespoke models complemented the experimental results, in that the findings supported the idea of uniform evolution for all pore-throat entrance sizes and provided a more robust modelling procedure which together enhanced the understanding of the mechanistic interpretations. Such findings contradict the current weight loss prediction models, but complement the working hypothesis formulated within EDF Energy graphite research group. Therefore, the experimental and modelled results will feature in a supportive document of proposed revisions to the EDF Energy Safety Case submitted to the Office for Nuclear Regulation.

Document Type


Publication Date




Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.