Dr Gareth Neighbour completed his PhD in Materials Science at the University of Bath and joined AEA Technology at their Windscale plant from 1992 for four years where, amongst other things, he managed the integration of project safety, QA and engineering management systems. He then moved to University of Bath where he formed and then led the internationally recognised Bath Nuclear Materials Group. Dr Gareth Neighbour joined the University of Hull in February 2001 as a lecturer in Engineering. A significant proportion of his research is interdisciplinary in nature. His interest ranges from integrated management systems to fracture mechanics and fundamental structure-property relationships of engineering materials. With over 64 publications and a significant number of classified reports published by nuclear industry organisations, he is often invited to collaborate with national and international organisations and assist with organising conferences and committees.

Modelling and Measuring Reactor Core Graphite Properties and Performance

By Gareth B. Neighbour

The Royal Society of Chemistry

Copyright © 2013 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-1-84973-390-8

Contents

Part A – Mechanistic,
Towards a Structural Basis to the Physical Properties of Irraidated Polycrystalline Nuclear Graphite Brian Rand, 2,
Property Changes of Polycrystalline Graphite due to Neutron Irradiation – A Critical Assessment after 70 Years of Research Gerd Haag, 9,
Evaluation of Instantaneous Coefficient of Thermal Expansion of PGA Graphite Martin P. Metcalfe, Nassia Tzelepi and John F.B. Payne, 17,
A Review of Current Finite Element Models for Irradiation Creep and Failure of Nuclear Graphite David W. J. Tanner, Adib A. Becker, Wei Sun and Tom H. Hyde, 23,
Quasi-Brittle Fracture Concepts to Improve Structural Integrity Assessments of High Oxidation Weight Loss Graphite Components Andrew D. Hodgkins, Mahmoud Mostafavi, Peter E. J. Flewitt, Malcolm Wootton and Robert Moskovic, 31,
Development of an Experimental and Simulation Process to Determine the End of Life Radionuclide Inventory of UK Irradiated Graphite Waste with a View to Long Term Disposal Greg Black, Abbie N. Jones and Barry Marsden, 39,
The Oxidation of A3-3 Matrix Material in a CO2 Atmosphere in Support of a Nuclear Battery Type Reactor Design Joel Turner, Abbie N. Jones, Anthony J. Wickham, Marc Schmidt and Tim J. Abram, 44,
Part B – Empirical,
Theoretical Investigations into Sample Size Effects on Ultrasonic Measurements of Elastic Moduli John F.B. Payne and Nassia Tzelepi, 52,
Mechanical Property Measurements on AGR Core Graphite Using Electronic Speckle Pattern Interferometry Paul Ramsay, 61,
Quantitative Microstructure Characterisation of Advanced Carbons Using Image Analysis Theerapatt Manuwong, Gary D. Kipling, and Gareth B. Neighbour, 69,
Revised Method For Graphite Weight Loss Prediction Jonathan E. Coote and Neil S. Headings, 76,
On Large Scale Implicit Multibody Contact Dynamics Modeling with SOLFEC Tomasz Koziara and Nenad Bicanic, 84,
Shaking Table Experimental Programme Luiza Dihoru, Adam Crewe, Colin Taylor and Tim Horgan, 91,
A Novel Approach for Evaluating Surface Deposits on the Channel Wall Trepanned Graphite Samples Peter J. Heard, Malcolm R. Wootton and Peter E. J. Flewitt, 99,
3H and 14C Release in UK Nuclear Graphite by Chemical Treatment Lorraine McDermott, Abbie N. Jones and Barry J. Marsden, 108,
Part C – Statistical,
Statistical Modelling of the Ageing of Graphite Cores Philip R. Maul, 116,
Probability Modelling of Density Variability of Radiolytically Oxidised Graphite Robert Moskovic, 124,
Statistical Analysis of Bore Cracking in AGRs Emma Tan, Kevin McNally and Nick Warren, 132,
Statistical Modelling of Graphite Properties Peter C. Robinson and Philip R. Maul, 140,
Part D – Plant,
The Role of Data Visualisation in Core Inspection Decision Making Claire E. Watson, Peter C. Robinson and Philip R. Maul, 149,
Non-Linear Seismic Assessment of AGRs: Incorporation of Aged Graphite Material Properties Kieran Scully and David Knowles, 158,
Solid-Body Simulation of the Dynamic Response of an Array of Graphite Bricks Steve Brasier and Stuart Rogers, 166,
Intelligent Graphite Core Condition Monitoring Christopher J. Wallace, Graeme M. West, Stephen D. J. McArthur and Michael Coghlan, 175,
Decomposition of Refuelling Signals to Estimate Channel Bore Profiles Graeme M. West, Christopher J. Wallace, Stephen D. J. McArthur and Michael Coghlan, 183,
The LEWIS Method for Channel Functionality Assessment Richard J. Crawford, 192,
Part E – Open Discussion,
Open Discussion Anthony J. Wickham, 202,
Author Index, 212,
Subject Index, 213,

CHAPTER 1

Towards a Structural Basis to the Physical Properties of Irradiated Polycrystalline Nuclear Graphite

Brian Rand

Institute for Materials Research, University of Leeds and Institute of Applied Materials (SARChl Chair of Carbon Technology and Materials), University of Pretoria, RSA

Email: b.ca.rand@btinternet.com

Abstract

The need for mechanistic understanding of structure-property relationships for irradiated polycrystalline nuclear graphite is considered. Appropriate structural characteristics for polycrystalline graphite at all of the relevant dimensional scales are discussed and some aspects of the effects of irradiation damage on them are considered. It is suggested that the relationship between structural characteristics and physical properties be determined at each dimensional scale and used in multi-scale modelling of specific properties at the macro-scale at which they are actually measured.

Keywords

Graphite, structure, properties

INTRODUCTION

Although significant progress was made in the periods when graphite core reactors were being developed and constructed, the following period has until recently been one of limited research activity due to the political perception of nuclear safety issues. Yet this very period is one when advanced material characterisation techniques have made tremendous strides along with the development of structural modelling techniques. Now that there is renewed interest in developing an improved understanding of the structure-property relationships for polycrystalline graphite, some of the earlier concepts inevitably will be re-evaluated. Currently, many of the material models in use to predict the changes in physical properties of nuclear graphite resulting from fast neutron irradiation and radiolytic oxidation are empirically based and their implementation is limited to the range of fluence and oxidative weight loss for which there is a validated database, hence the need for better mechanistic understanding of the structural basis of the properties and the development of mechanistically based models for the different properties. This paper presents a personal view of what may be required to develop such an approach.

STRUCTURE-PROPERTY RELATIONSHIPS

For the UK situation, perhaps the most important mechanistic issues for polycrystalline material are:-

• The mechanism of dimensional change, since the differential changes control internal stresses

• Irradiation creep which leads to the alleviation of such stresses;

• The structural changes due to fast neutron irradiation and to radiolytic oxidation;

• Structural dependence of physical properties.

Given this level of understanding, it should be possible to understand the effects of irradiation damage and oxidation on the physical properties of interest. Such a task, however, is not easy due to the structural complexity of polycrystalline graphite. Figure 1 shows the obvious relationship between structural characteristics and physical properties. When the former are changed by irradiation and oxidation, the physical properties are changed. However, if we cannot clearly define what these structural characteristics are and/or do not know how they are changed by irradiation then the modified physical properties cannot be predicted. Consequently, because we do not currently have this level of understanding and because it has been relatively easy to determine the changes in physical properties the tendency has been to use the physical property database developed for prediction of core component behaviour and then for mechanistic understanding to attempt to deduce how the structure has been modified, as shown by the grey arrows in Figure 1.

Such an approach is inevitably limited in its scope. The essential database of properties is created, but its ability to be extrapolated to fluences much beyond the extent of the database is severely limited. The level of understanding has also been supplemented by the study of materials approximating to the single crystal state and whilst such studies have invaluable, they also are limited in providing insight only into the behaviour of the material at the crystallographic level as explained in more detail below.

STRUCTURE OF POLYCRYSTALLINE GRAPHITE

It is important first of all to define the important structural characteristics that will determine the relevant physical properties. This is not straight forward as there are important structural features at all dimensional scales, ranging from nano- to micro- to macroscopic that may in principle contribute to a particular physical property, the relative importance of each being different for each property. It is relatively easy to describe these qualitatively as below but more problematic to define quantitative characteristics that can be measured and related directly to specific physic-chemical properties, yet such an approach is required.

• At the nano-scale there are the regions showing coherent x-ray diffraction, often known as ‘crystallites’, defined by the coherence lengths in the two principal crystallographic directions, La and Lc and the interlayer spacing, d002. There may also be nano-scale fissures associated with (separating) such regions arising perhaps from thermal contraction stresses during cool down from graphitization temperatures. The behaviour of this fundamental unit is usually understood from studies of natural flake graphites and highly oriented pyrolytic graphites (HOPG) (Kelly, 1981). A range of basal and non-basal dislocations has been identified and the effects of irradiation damage at typical reactor temperatures in creating dislocation loops was studied in some detail (Kelly, 1981; Thrower, 1969). The dimensional changes accompanying irradiation have been documented at various temperatures (Kelly, 1981). However, the existence of such dislocations in nuclear graphites has not been well documented, there being limited recorded observations of dislocation for such material (Shtrombakh et al., 1995). Their role is also not so clear due to the more limited extent of the coherent regions and the uncertain nature of the boundaries between them. Modern advanced characterisation techniques should in future enable a better understanding of the defect/dislocation character of irradiated nuclear graphites. Quantitative characteristics might be:

* La from x-ray diffraction and Raman spectroscopy.

* Lc and d002 from x-ray diffraction.

* Dislocation loop density.

* sp2 character.

• These coherence regions are not isolated but are connected to form domains of microscopic dimensions, extending over tens to hundreds of microns and visible in the optical microscope under crossed polars. Figure 2 shows a scanning electron micrograph (SEM) of such a region. A highly ordered ‘needle’ particle is shown in the bottom right hand comer and there are twisted layer packets or domains in the remainder of the view. In some publications, these domains (or regions within the domains, Shtrombakh et al. (1995)) have been taken to represent the crystallites and there is some justification for this as discussed below. The domains may be bent and twisted and within them there may be additional lamellar fissures, again arising from thermal contraction stresses, but also from the shrinkage arising from the densification of the graphite during manufacture as the true density increases from about 1.4 at the mesophase state where the lamelliform structure is developed through about 1.9 on ‘baking’ to 2.25 g cm-3 at the graphite stage. These fissures are understood to affect the thermal expansion behaviour and the dimensional change by partly accommodating the c-axis expansion and the c-axis dimensional change. Those fissures arising from thermal contraction (the so-called ‘Mrozowski’ cracks) may close progressively on heating whilst those arising from structural densification may not. Both, however, will contribute to the dimensional change. Thus, there is a local (micro-scale) preferred orientation of the coherence regions and the associated fissures that needs to be understood. The fundamental understanding that we have for the behaviour of ‘crystallites’ (derived from the study of HOPG for example) may more precisely relate to regions of connected coherence lengths with a preferred orientation lying within the larger domains and between the lamellar fissures and this will be the basic driving force for dimensional change. The behaviour of the domain, however, will be different due to the moderating effect of the fissures on expansion behaviour. The physical properties locally within a domain will also be different from those locally within the coherence region and of the connected coherence region. Quantitative parameters include:

* Domain dimensions in different directions (shape).

* Preferred orientation of layer planes within the domains (extent of wrinkling).

* Intra-domain porosity (fissures).

* Orientation of porosity.

* Local physical properties.

There will be a connection between the structural parameters at the nano-scale and some of the domain parameters. For example, the local preferred orientation within a domain will be changed as the coherence lengths are reduced by irradiation damage.

• At the macro-scale, there is the arrangement of domains. At some dimensional scale, there will be a preferred orientation of the domains, perhaps within the boundaries of a filler grain, but on a larger scale of size the orientation of domains may be isotropic or there may be a preferred orientation as in PGA graphite. At the macro-scale, we have to consider the filler and binder regions, which will show different domain textures although not necessarily vastly different crystallographic structure ( i.e., in terms of the coherently diffracting regions). It is certainly not true that the binder regions are amorphous as has occasionally been claimed. Quantitative parameters could comprise:

* Filler grain/binder proportions.

* Filler grain size/shape distribution.

* Preferred orientation of anisometric filler grains.

* Preferred orientation of domains within filler grains.

* Domain dimensions within filler and binder.

* Intra-porosity of filler grains.

* Inter filler porosity.

So, whilst much has been made of the changes at the so-called ‘crystallite’ level, which may refer to coherence lengths or to domains, and it is surely the case that the driving force for structural rearrangement derives at this scale, in determining the effects on a macroscopic property some model characteristic structural unit at the macroscopic scale, which is a composite of the above structural features, will be the key to modeling a particular physical property.

EFFECT OF IRRADIATION ON STRUCTURE

The standard model of the effects of fast neutron irradiation on the structure (see, for example, Kelly (1981)) considers that interstitial carbon atoms form interplanar discs (dislocation loops) which lead to expansion of the crystal in the c-direction whilst vacancies created condense into lines leading to contraction in the a-direction. However, this view is currently questioned by Heggie and co-workers (2011), who have proposed a so-called ‘ruck and tuck’ model of damage and dimensional change operative at temperatures typical of reactor operation. The model involves the shearing of basal planes by dislocation movement. The validity of this new approach remains to be established although it has many attractive features. X-ray and transmission electron microscopy studies (TEM) have established that the coherent lengths show a drastic reduction after exposure to significant fluences with the TEM images showing clear ‘wrinkling’ of the layer planes (Asthana et al., 2005) which resembles the structures in the coke at the carbon stage prior to complete graphitization of the material, i.e. the material becomes almost turbostratic (Tanabe and Muto, 1999). This raises the questions, “to what extent can irradiation damage be regarded as a reverse of the graphitization process” and “how valid is the ‘ruck and tuck’ model to graphitization itself”? The break up of the crystallographic structure creates many more boundaries between the coherently diffracting regions and, within the domains discussed above, there should be a corresponding change in the local preferred orientation. This structural change to a turbostratic state and the creation of these boundaries raises the question of the extent to which dislocation movement is relevant to the behaviour of irradiated graphite. A current model of irradiation creep by Kelly and Foreman (1974), for example, proposes that the secondary creep process is controlled by the pinning and unpinning of dislocations by neutron bombardment, but their paper acknowledges that “A dislocation cannot be defined in a turbo-stratic structure.” Thus, the role of dislocations in determining the properties of graphite irradiated to high fluence requires clarification.

It is commonly accepted that there are changes to the network of fissures and more equi-axed pores that comprise the porosity in the material. This is well documented by the study of Engle (1971) who used a combination of pycnometric techniques to quantitatively assess the changes to various types of pores in two graphites through the region at which dimensional change turnaround occurs and beyond. He was able to show how pores between crystallite clusters were closed by the c-axis expansion and other pores opened up between crystallite clusters. Beyond turnaround in the macroscopic expansion phase, new pores were also formed between filler particles. However, the body of evidence for the structural changes through the whole extent of fluence likely to be experienced during reactor service is sparse and needs to be improved with the application of the modern methods available these days.

(Continues…)Excerpted from Modelling and Measuring Reactor Core Graphite Properties and Performance by Gareth B. Neighbour. Copyright © 2013 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.