Establish models for predicting physical properties of a nuclear graphite based on its experimentally observed microstructure

点击翻译按钮获取中文摘要

Until now, there is no complete model available to predict physical properties of nuclear graphite based on experimentally observed microstructure. On the other hand, for most key metallic and ceramic materials, the correlations between physical properties and microstructure have been reasonably well established, making component design and lifetime management much easier and less time-consuming. Engineers/scientists working on nuclear graphite, however, are still relying on extremely expensive and very time-consuming experimental work to acquire most needed physical properties through accelerated neutron irradiation in a material testing reactor (MTR). One of the key reasons for such a scenario is the incredibly complicated microstructure of nuclear graphite, particularly across dimension scale from nano-meter to micro-meter. Over the past several years, the Loughborough Nuclear Graphite Group (LNGG) has been investigating the microstructure and related structural characteristics of nuclear graphite. Among the outcomes, "crazy paving" has been experimentally and theoretically evidenced as the key fundamental structural units. Such units are the primary building blocks for a nuclear graphite, and the results have been published recently.Based on the progress and preliminary trial in modelling, we hence propose this research for a 3 year PhD studentship aiming to achieve the goal. a) Establish models based on experimentally observed microstructure We believe "crazy paving" should be the basic building block in nuclear graphite. Research in this task will experimentally establish how such building blocks extended in 3D dimension inside fillers and binders in synthetic graphite, including inside key characteristic features as spherical QI particles and fillers. HRTEM/STEM will be used to illustrate the crazy paving structure of samples made from specifically interested regions in graphite with FIB microscopy. The methodology has been reasonably well established through research in UNIGRAF, and a PhD researcher is expected to get hands on from the start of the project. At the same time, Raman spectroscopy data from these regions will be collated by the measurements from HRTEM/STEM, aiming to establish an efficient and non-destructive tool to quantify the crazy paving structure. Measurements from Raman spectroscopy will be tested if they can be used as modelling inputs to predict physical properties. b) Develop appropriate modelling tool kits to predict physical properties based on these models we will use atomistic modelling based on a combination of ab initio and classical techniques to investigate prediction tools. The research will also use in-house developed computer code and models to investigate extended time scales normally outside the range of application of these methods. By now, the Loughborough research team has 2 PhD and 2 RA researchers who are working on different atomistic modelling. This PhD researcher will be part of the team but with a clear focus on predicting properties. The modelling will start from simple "crazy paving" structure, and gradually move into more complicated structure built up from "crazy paving" building blocks based on measurements from task a). The prediction will be able to extend to a geometric dimension that nano/micro-scale measurements can be possible. c) Validation of modelling predictionDue to the complexity of graphite structure, it might be too challenge by using the routine physical property measurements to directly validate predictions from tools based on atomistic modelling. However, if property measurement can be done at nanometer or submicron levels, such modelling prediction can be validated. Hence, in this research, we will directly measure key physical properties inside electronic microscopes. We will first focus on measurement of CTE on a heating stage installed in TEM with specimen size dimensions ranging from nanometers to sub-microns. These samples