using intercalation to simulate irradiation damage of

222
Using Intercalation to Simulate Irradiation Damage of Nuclear Graphite Lewis Luyken 2012 A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Engineering and Physical Sciences. School of Mechanical, Aerospace and Civil Engineering The University of Manchester

Upload: others

Post on 02-Apr-2022

10 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Using Intercalation to Simulate Irradiation Damage of

Using Intercalation to Simulate Irradiation

Damage of Nuclear Graphite

Lewis Luyken

2012

A thesis submitted to the University of Manchester for the

degree of Doctor of Philosophy in the Faculty of Engineering

and Physical Sciences.

School of Mechanical, Aerospace and Civil Engineering

The University of Manchester

Page 2: Using Intercalation to Simulate Irradiation Damage of

2

Acknowledgements

I owe a huge thanks to Professor Barry Marsden for giving me the

opportunity to carry out this research. I never imagined that a blanket

email to the professors in the school searching for a supervisor for a

half baked undergraduate project would lead on to this. His support

throughout the project has been invaluable.

Secondly I am hugely indebted to Dr Abbie Jones, Dr Marc Schmidt, Dr

Med Benyezzar, Professor James Marrow, Professor Paul Mummery

and Mr Rob Stringer. Without their advice and assistance I could never

have completed this work. I also owe a big thanks to the members of

the nuclear graphite research group who have made the project an

enjoyable experience even while in the depths of research doom.

Finally Rosie Luyken whose patience and support has been invaluable, I

am so lucky to have her in my life. Most importantly I will now be able

to spend much more time causing havoc with Seumas.

Page 3: Using Intercalation to Simulate Irradiation Damage of

3

Declaration

No portion of the work presented in the thesis has been submitted in support of an

application for another degree or qualification of this or any other university or

other institute of learning.

The author of this thesis (including any appendices and/ or schedules to this thesis)

owns certain copyright related rights in it (the “copyright” and s/he has given the

University of Manchester certain rights to use such copyright including for

administrative proposes.

Copies of this thesis, wither in full or in extracts and whether in hard or electronic

copy, may be made only in accordance with the copyright, Designs and Patents Act

1988 (as amended) and regulations issued under it or, where appropriate, in

accordance with licensing agreements which the university has from time to time.

This page must form part of any such copies made.

The ownership of certain Copyright, patents, designs, trade marks and other

intellectual property ( the “intellectual property”) and any reproduction of copy

right works in the thesis, for example graphs and tables (“Reproductions”), which

may be described in this thesis, may not be owned by the author and may be

owned by third parties. Such intellectual Property and Reproductions cannot and

must not be made available to use without the prior written permission of the

owner(s) of the relevant Intellectual Property and/ or Reproductions.

Further information on the conditions under which disclosure, publication and

commercialisation of this thesis, the copyright and intellectual property and/ or

Reproductions described in it may take place is available in the University IP Policy

(see https://documnets.manchester.ac.uk/DocuInfo.aspx?DocID=487) , in any

relevant thesis restriction declarations deposited in the University Library, The

University Library’s regulations (see

http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s

policy on Presentation of Theses.

Page 4: Using Intercalation to Simulate Irradiation Damage of

4

Abstract

This thesis investigates the use of bromine intercalation of graphite as a

method to simulate and investigate irradiation damage. In particular

this study investigates the effects of intercalation on dimensional

change on the macro and micro scales and how these changes combine

to affect Young’s modulus.

Highly Orientated Pyrolytic Graphite has been used to gather data as a

close approximation to single crystal graphite. Three different grades of

polycrystalline nuclear graphite have been used to investigate the

effect of different microstructure on intercalation and subsequent

property changes. The graphites have been characterized by optical

microscopy, pycnometry and x-ray powder diffraction and texture

measurements. A number of bespoke rigs were designed and

manufactured to carry out sorption, tomography and laser vibrometry

experiments.

The results indicate that the rate of dimensional change for

polycrystalline graphites is significantly lower than for single crystal

graphites. Modelling of dimensional change suggests that the

difference in expansion is due to closure of porosity. Closer

investigation of the dimensional change within the microstructure

shows that the majority of the dimensional change is driven by

expansion of filler particles.

The young’s modulus results show an initial increase in modulus

followed by a decrease, which corresponds with empirical evidence for

irradiated samples. It is postulated that the initial increase in modulus is

Page 5: Using Intercalation to Simulate Irradiation Damage of

5

due to crystal expansion and that the subsequent decrease is due to

crack growth. After experimentation some samples show significant

cracking which would appear to support this assertion.

Page 6: Using Intercalation to Simulate Irradiation Damage of

6

Contents

Acknowledgements ........................................................................................... 2

Declaration ........................................................................................................ 3

Abstract ............................................................................................................. 4

Contents ............................................................................................................ 6

List of Figures ................................................................................................... 10

List of Tables .................................................................................................... 17

Nomenclature .................................................................................................. 18

Abbreviations .................................................................................................. 20

1. INTRODUCTION ....................................................................................... 22

1.1. Development of Graphite Reactors ..................................................... 23

1.1.1. Chicago Pile Zero ...................................................................................... 23

1.1.2. Magnox Reactors ...................................................................................... 23

1.1.3. Advanced Gas Cooled Reactor .................................................................. 24

1.1.4. Very High Temperature Reactors .............................................................. 25

1.2. An Introduction to Nuclear Graphite ................................................... 26

1.2.1. Cross Section ............................................................................................ 27

1.2.2. Structure and manufacture....................................................................... 28

1.2.3. Manufacture ............................................................................................ 34

1.3. Irradiation Damage .............................................................................. 36

1.3.1. Atomic Scale Damage ............................................................................... 36

1.3.2. Dimensional Change ................................................................................. 39

1.3.3. Young’s Modulus ...................................................................................... 47

1.4. Scientific Proposal ............................................................................... 51

Page 7: Using Intercalation to Simulate Irradiation Damage of

7

1.5. Summary ............................................................................................. 52

2. INTERCALATION ...................................................................................... 53

2.1. Intercalation Compounds .................................................................... 53

2.2. Graphite Intercalation Compounds ..................................................... 55

2.2.1. Intercalation Methods .............................................................................. 55

2.2.2. Structure of Graphite Intercalation Compounds ....................................... 57

2.2.3. The Intercalation Reaction ........................................................................ 61

2.3. Property Changes ................................................................................ 68

2.3.1. Dimensional Changes ............................................................................... 68

2.3.2. Young’s Modulus ...................................................................................... 69

2.4. Summary ............................................................................................. 71

3. MATERIALS AND METHODS .................................................................... 73

3.1. Experimental Techniques .................................................................... 73

3.1.1. McBain Spring Balance ............................................................................. 73

3.1.2. X-ray Computed Tomography ................................................................... 74

3.1.3. Laser Ultrasonics ...................................................................................... 80

3.1.4. X-ray Diffraction ....................................................................................... 85

3.2. Experimental Rigs ................................................................................ 90

3.2.1. Design of experimental rig to measure dimensional change of bulk material

................................................................................................................. 90

3.2.2. Design of experimental rig to measure microstructural dimensional change

................................................................................................................. 91

3.2.3. Design of experimental rig to measure Young’s modulus change of

brominated graphite ................................................................................ 95

3.3. Conclusions .......................................................................................... 99

4. CHARACTERISATION OF NUCLEAR GRAPHITE ....................................... 100

Page 8: Using Intercalation to Simulate Irradiation Damage of

8

4.1. Characterisation Results .................................................................... 100

4.1.1. Polarised optical microscopy .................................................................. 100

4.1.2. Pycnometry ............................................................................................ 102

4.1.3. Powder Diffraction ................................................................................. 103

4.1.4. Textural Analysis..................................................................................... 107

4.2. Graphite Grades ................................................................................ 113

4.2.1. Pile Grade A graphite .............................................................................. 113

4.2.2. Pechiney Graphite .................................................................................. 114

4.2.3. Gilsocarbon ............................................................................................ 114

4.3. Conclusions ........................................................................................ 115

5. DIMENSIONAL CHANGE ........................................................................ 116

5.1. Dimensional Change of Nuclear Grade Graphite by Bromine

Intercalation ...................................................................................... 118

5.1.1. Calibration of the spring for sorption balance ......................................... 119

5.1.2. Dimensional change experiments ........................................................... 121

5.1.3. Dimensional Change of Highly Orientated Pyrolytic Graphite .................. 122

5.1.4. Dimensional Change of Polycrystalline Graphite ..................................... 132

5.1.5. Analysis of Dimensional Change ............................................................. 137

5.2. Microstructural Experiment .............................................................. 156

5.2.1. Two dimensional Analysis ....................................................................... 158

5.3. Conclusions ........................................................................................ 172

6. YOUNG’S MODULUS.............................................................................. 174

6.1. Measuring optimal input energy for Young’s modulus measurements

by laser impact excitation ................................................................. 176

6.1.1. Verification of Laser Impact Excitation as a suitable modification of ASTM

C769 ....................................................................................................... 179

6.2. Change in modulus of brominated graphite ...................................... 184

Page 9: Using Intercalation to Simulate Irradiation Damage of

9

6.2.1. Change in density of brominated graphite .............................................. 185

6.3. Modulus changes of brominated Graphite ........................................ 189

6.3.1. HOPG ..................................................................................................... 189

6.3.2. Modulus changes in brominated polycrystalline graphites ...................... 195

6.3.3. Possible sources of Errors ....................................................................... 204

6.4. Comparison of modulus changes due to irradiation and bromination

.......................................................................................................... 205

6.5. Conclusions ........................................................................................ 208

7. CONCLUSIONS AND FURTHER WORK .................................................... 210

7.1. Conclusions ........................................................................................ 210

7.2. Future Work ...................................................................................... 212

8. REFERENCES .......................................................................................... 214

Page 10: Using Intercalation to Simulate Irradiation Damage of

10

List of Figures Figure 1.1 Magnox Schematic[2] ....................................................................................... 24

Figure 1.2 AGR Schematic [3] ............................................................................................ 25

Figure 1.3 Very High Temperature Reactor [6] ................................................................... 26

Figure 1.4 Graphite Structure[11] 30

Figure 1.5 Atomic structure of graphite crystal[10] ............................................................ 30

Figure 1.6 Edge Dislocation [14] ....................................................................................... 31

Figure 1.7 Screw Dislocation[15] ....................................................................................... 31

Figure 1.8 TEM of Highly Orientated Pyrolytic Graphite (Image courtesy of Keyen Wen) ... 32

Figure 1.9 Gilsocarbon Microstructure (Image Abbie Jones) .............................................. 33

Figure 1.10 Binder Matrix Crystallite Structure (Image Abbie Jones) .................................. 33

Figure 1.11 Graphite manufacturing process [18] .............................................................. 34

Figure 1.12 Displacement Cascade .................................................................................... 36

Figure 1.13 Growth of interstitial loops[22] ....................................................................... 37

Figure 1.14 Frenkel Pair[1] ................................................................................................ 37

Figure 1.15 Interstitial loops created by irradiation at 1350°C to 11.7 x 1020 neutrons cm-2

[27] ................................................................................................................................... 38

Figure 1.16 Effect of dose and irradiation temperatures on dimensional change of HOPG in

the c-axis[25] .................................................................................................................... 40

Figure 1.17 Effect of dose and irradiation temperature on the dimensional change of HOPG

in the a-axis.[25] ............................................................................................................... 40

Figure 1.18 PGA Dimensional Change at low Temperature[35] .......................................... 41

Figure 1.19 PGA Dimensional Changes of PGA at High Temperatures Perpendicular to

Extrusion[36] .................................................................................................................... 42

Page 11: Using Intercalation to Simulate Irradiation Damage of

11

Figure 1.20 Dimensional Changes at High Temperatures Parallel to Extrusion[36] ............. 42

Figure 1.21 Dimensional changes of Gilsocarbon due to irradiation damage[32] ............... 43

Figure 1.22 Schematic describing the with grain and against grain directions .................... 46

Figure 1.23 Nuclear graphite stress strain curve[50] .......................................................... 48

Figure 1.24 Orientation of unit vectors in relation to graphite layer planes........................ 49

Figure 1.25 Change in Gilsocarbon Young's modulus due to irradiation[32, 36] ................. 51

Figure 2.1 Bromine Intercalation Apparatus for weight uptake measurements[58] ........... 56

Figure 2.2 Bromine Intercalation with control of intercalant partial pressure for XRD

measurements[76] ............................................................................................................ 57

Figure 2.3 Schematic of increasing intercalation stages ..................................................... 58

Figure 2.4 TEM Image of CuCl2 (light regions) intercalated graphite (dark regions) [79] ..... 58

Figure 2.5 Distance between graphite layer planes for different intercalation species [58] 59

Figure 2.6 Graphite basal planes (black) with intercalant (white) a) stage 3 b) stage 2 ....... 59

Figure 2.7 Bromine positions at low concentrations calculated by DFT [66] ....................... 60

Figure 2.8 Unit Cell C8Br found by Neutron Diffraction [86] ............................................... 60

Figure 2.9 Intercalate clusters [24] .................................................................................... 61

Figure 2.10 Interstitial clusters caused by fast irradiation[27] ............................................ 61

Figure 2.11 Phase change induced by change in intercalant pressure. [76] ........................ 62

Figure 2.12 Intercalation Rate through sample thickness [91] ............................................ 63

Figure 2.13 Stage two and stage three Daumas Herold Domains[94] ................................. 64

Figure 2.14 Effect of crystal perfection on bromine uptake [101] ...................................... 67

Figure 2.15 Change in elastic constants of lithiated graphite[109] ..................................... 70

Page 12: Using Intercalation to Simulate Irradiation Damage of

12

Figure 3.1 Schematic of McBain Spring Balance[113] ......................................................... 74

Figure 3.2 Tomography setup ............................................................................................ 75

Figure 3.3 Effect of sample composition on x-ray shadow ................................................. 77

Figure 3.4 Schematic of ASTM C769 experimental setup[129] ........................................... 81

Figure 3.5 Schematic of ultrasonic wave generation by laser [131] .................................... 83

Figure 3.6 Michellson interferometer ................................................................................ 85

Figure 3.7 Philips X’pert modular diffractometer ............................................................... 86

Figure 3.8 Schematic representation of diffraction by a crystal[36, 136] ............................ 86

Figure 3.9 Schematic of hkl planes[18] .............................................................................. 87

Figure 3.10 Micrometrics pycnometer ............................................................................... 90

Figure 3.11 Dimensional Change Rig .................................................................................. 91

Figure 3.12 X-ray attenuation coefficients of candidate rig materials[144] ........................ 93

Figure 3.13 Tomography sample ....................................................................................... 94

Figure 3.14 Experimental rig for tomographic scans of bromine intercalated graphite ....... 95

Figure 3.15 Reflectivity of argon coated glass[146] ............................................................ 97

Figure 3.16 Reflectivity of VIS-NIR coated glass[146] ......................................................... 97

Figure 3.17 Exploded diagram of cell to measure Young's modulus of brominated graphite

......................................................................................................................................... 98

Figure 3.18 Sample holder for polycrystalline graphite ...................................................... 99

Figure 4.1 Optical mircographs of nuclear grade graphite ................................................ 102

Figure 4.2 XRD powder diffraction spectra of PGA ........................................................... 105

Figure 4.3 XRD powder diffraction spectra of Pechinay Graphite ..................................... 105

Page 13: Using Intercalation to Simulate Irradiation Damage of

13

Figure 4.4 XRD powder diffraction spectra of Gilsocarbon ............................................... 106

Figure 4.5 Image detailing the Full Width at Half Maximum[148] .................................... 107

Figure 4.6 PGA pole figure ............................................................................................... 110

Figure 4.7 Pechiney pole figure ....................................................................................... 111

Figure 4.8 Gilsocarbon pole figure ................................................................................... 112

Figure 5.1 Brick deformations caused by non uniform flux profiles .................................. 116

Figure 5.2 McBain Sorption Balance[58] .......................................................................... 119

Figure 5.3 Calibration of spring for weighing polycrystalline samples............................... 120

Figure 5.4 Calibration of spring for weighing HOPG samples ............................................ 120

Figure 5.5 Schematic of McBain Spring Balance ............................................................... 121

Figure 5.6 Schematic of crystal arrangement in HOPG 126

Figure 5.7 Schematic of HOPG sample with orientation of axes shown ............................ 124

Figure 5.8 Dimensional change of bromine intercalated HOPG 130

Figure 5.9 HOPG microstructure ...................................................................................... 128

Figure 5.10 Difference between theoretical and measured mass ..................................... 129

Figure 5.11 In plane expansion ........................................................................................ 130

Figure 5.12 Irradiation induced dimensional change of HOPG[33] ................................... 131

Figure 5.13 Dimensional change of extruded polycrystalline graphites ........................... 134

Figure 5.14 Dimensional change of PGA and Gilsocarbon ................................................ 135

Figure 5.15 Dimensional change due to irradiation[32] ................................................... 137

Figure 5.16 Porosity distribution base graphite[18] ......................................................... 139

Figure 5.17 XRD texture orientations .............................................................................. 142

Page 14: Using Intercalation to Simulate Irradiation Damage of

14

Figure 5.18 Theta derived for against grain calculation ................................................... 142

Figure 5.19 Theta derived for with grain calculation ........................................................ 144

Figure 5.20 Effect of crystal orientation on dimensional change against grain ................. 145

Figure 5.21 Effect of crystal orientation on dimensional change with grain ..................... 146

Figure 5.22 PGA bromination accommodation factors ..................................................... 149

Figure 5.23 Gilsocarbon bromination accommodation factors…………………………….............149

Figure 5.24 PGA thermal expansion accommodation factors[158] ................................. 150

Figure 5.25 PGA predicted dimensional change ............................................................... 151

Figure 5.26 Modelled dimensional change of Gilsocarbon ............................................... 152

Figure 5.27 Accommodation Volume .............................................................................. 153

Figure 5.28 Prediction of Irradiated Dimensional Change using intercalation

accommodation factors; simulation of low neutron dose ................................................ 155

Figure 5.29 Prediction of irradiated dimensional change using intercalation accommodation

factors; simulation of high neutron dose ......................................................................... 156

Figure 5.30 Bromination rig in position. Highlighted are a) X-ray source b) shutter c) rig and

d) camera ........................................................................................................................ 158

Figure 5.31 Digital image correlation performed on PGA radiographs .............................. 159

Figure 5.32 Dimensional change calculated from radiographs for PGA ............................ 161

Figure 5.33 Dimensional change calculated from radiographs for Gilsocarbon................. 162

Figure 5.34 Dimensional Change calculated by digital image correlation for Pechinay ..... 162

Figure 5.35 Bicubic filter .................................................................................................. 164

Figure 5.36 Microstructure of PGA tomography sample, before bromination ................. 165

Figure 5.37 Change in grayscale of brominated samples .................................................. 171

Page 15: Using Intercalation to Simulate Irradiation Damage of

15

Figure 6.1 Ultrasonic pulse measured with laser vibrometer ........................................... 177

Figure 6.2 Ultrasonic pulse measured with laser vibrometer ........................................... 178

Figure 6.3 Effect of Impact Energy on Measured Waveform ............................................ 179

Figure 6.4 ASTM C769 Experimental setup for piezo transducer measurement of Young's

modulus .......................................................................................................................... 180

Figure 6.5 Time of flight data for modulus measurement by piezo transducer ................. 181

Figure 6.6 Experimental setup for measurement of Young's Modulus by Laser Impact

Excitation and Laser vibrometry ...................................................................................... 182

Figure 6.7 Schematic of experimental setup for measurement of Young's Modulus by Laser

Impact Excitation and Laser vibrometry .......................................................................... 183

Figure 6.8 Comparison of sonic velocity measurements by Laser and Piezo Techniques .. 184

Figure 6.9 Change in density of HOPG ............................................................................. 188

Figure 6.10 Density change of polycrystalline graphites ................................................... 188

Figure 6.11 Laser confocal micrograph of debrominated PGA a) green scale bar 100µm area

of image b highlighted in green b) scale bar 40µm filler particle outlined in green ......... 189

Figure 6.12 Piezo setup for brominated samples. ............................................................ 190

Figure 6.13 Change in Modulus of Brominated HOPG ...................................................... 192

Figure 6.14 Change in Young's modulus of brominated Pile Grade A cut perpendicular to

extrusion (Against Grain) ................................................................................................. 198

Figure 6.15 Pile Grade A sample 1 cut against the grain (perpendicular to extrusion) after

bromination .................................................................................................................... 199

Figure 6.16 Pile Grade A sample 3 cut against the grain (perpendicular to extrusion) after

bromination .................................................................................................................... 199

Page 16: Using Intercalation to Simulate Irradiation Damage of

16

Figure 6.17 Change in Young's Modulus of Brominated Pile Grade A cut Parallel to Extrusion

....................................................................................................................................... 200

Figure 6.18 Pile Grade A Sample after bromination ......................................................... 201

Figure 6.19 Young's modulus changes in brominated Gilsocarbon ................................... 203

Figure 6.20 Laser confocal micrograph of debrominated Gilsocarbon showing reltivly little

microstructural cracking a) scale bar 100µm b) isometric image 1280 x 1280 µm ........... 203

Figure 6.21 Comparison of changes in Young's modulus in PGA parallel to extrusion due to

irradiation and bromination ............................................................................................ 206

Figure 6.22 Comparison of changes in Young's modulus in PGA parallel to extrusion due to

irradiation and bromination ............................................................................................ 207

Figure 6.23 Comparison of changes in Young's modulus in Gilsocarbon due to irradiation

and bromination ............................................................................................................. 207

Page 17: Using Intercalation to Simulate Irradiation Damage of

17

List of Tables Table 1.1 Moderator Cross Sections ............................................................................ 28

Table 1.2 Voight Notations .......................................................................................... 28

Table 3.1 Theoretical loss of x-ray intensity due to rig ................................................. 93

Table 3.2 Microscope objectives available on TOMCAT beamline[143] ........................ 94

Table 4.1 Powder diffraction data ............................................................................. 106

Table 5.1 Properties of ZYH HOPG ............................................................................. 124

Table 5.2 Tabulated Results of HOPG dimensional change ........................................ 126

Table 5.3 Analysis of c axis expansion ....................................................................... 128

Table 6.1 Fractional Uncertainties in Young’s modulus measurement of HOPG ......... 195

Page 18: Using Intercalation to Simulate Irradiation Damage of

18

Nomenclature

α accommodation factor in a

β crystal CTE

γ accommodation factor in c

δ dose

η Poisson’s ratio

λ wavelength

μ x-ray absorption co-efficient

ξ unit vector

ρ density

φ tilt angle

φa wave polarization

ψ azimuth angle

B Full width at half maximum

C elastic constant

d interatomic spacing

D.C. Dimensional change

E Young’s modulus

g irradiation growth of a crystal

I intensity

k spring constant

K Bacon anisotropy factor

l length

m mass

P Pressure

T Temperature

Tαβ Stress vector

Page 19: Using Intercalation to Simulate Irradiation Damage of

19

uστ Strain vector

Page 20: Using Intercalation to Simulate Irradiation Damage of

20

Abbreviations

AG Against Grain

AGL Anglo Great Lakes

AGR Advanced Gas cooled Reactor

ASTM American Society for Testing and Materials

BAEL British Acheson Electrodes Ltd

CCD Charged Couple Device

CTE Co-efficient of Thermal Expansion

DFT Density Functional Theory

DIC Digital Image Correlation

DYM Dynamic Young’s Modulus

EMAT Electro Magnetic Acoustic Transducer

FWHM Full Width at Half Maximum

HOPG Highly Orientated Pyrolytic Graphite

HTR High Temperature Reactor

MBA Modified Bronnikov Algorithm

NDT Non-Destructive Testing

PGA Pile Grade A graphite

PTFE Polytetrafluoroethylene

SLS Swiss Light Source

STM Scanning Tunnel Microscope

TEM Transmission Electron Microscope

Page 21: Using Intercalation to Simulate Irradiation Damage of

21

THTR Thorium High Temperature Reactor

VHTR Very High Temperature Reactor

WG With Grain

XCT X-ray Computed Tomography

XRD X-Ray Diffraction

Page 22: Using Intercalation to Simulate Irradiation Damage of

22

1. Introduction

This project aims to better understand the underlying principles of

material property changes induced in graphitic components within a

nuclear reactor during operation. This will be achieved using

intercalation as a technique to simulate irradiation damage to

understand material property changes.

This work will be of use to scientists and engineers within the nuclear

industry as a deeper understanding of property changes will allow

operators to improve safety cases for continued reactor operation and

allow manufacturers to improve the design of graphite for future

reactor designs. Intercalation is an area of interest across many areas of

science in particular this research may be of interest to battery

manufactures where intercalation of graphite is also a life limiting

feature.

There are many different reactor designs in use and graphite is used as

a critical component in many of them. Graphite has a unique

combination of properties in that it absorbs few neutrons whilst

remaining structurally and chemically stable at high temperatures.

Reactor designers use these properties to achieve a sustained nuclear

reaction with components that are structurally stable across a large

range of temperatures and neutron doses. As a result there is a huge

body of research dedicated to the performance of graphite under fast

neutron irradiation. Irradiation induced material property changes to

Page 23: Using Intercalation to Simulate Irradiation Damage of

23

bulk graphite are generally attributed to changes in the atomic

structure.

1.1. Development of Graphite Reactors

1.1.1. Chicago Pile Zero

The first self sustaining artificial nuclear fission reaction was achieved

for 28 minutes by a team lead by Enrico Fermi in December 1942. A

reactor known as Chicago Pile 1, was built in a squash court on the

University of Chicago campus. The reactor was constructed from a pile

of graphite blocks interspersed with 40 tons of uranium oxide and 6

tons of uranium metal for fuel. Compared to today’s designs it was very

low tech with ambient air cooling and a man with an axe to cut a rope

attached to a boron rod which would shut the reaction down in an

emergency [2].

1.1.2. Magnox Reactors

The next significant advance for the civilian use of nuclear power

occurred in the UK with the construction of the world’s first commercial

reactors at Calder Hall. Calder hall comprises four graphite moderated

reactors with significant technological advancements over pile type

reactors. The fuel and reactor are contained within a pressure vessel

cooled by CO2 and operated at around 390°C. This provided an output

of 200 MWe[3]. The design became known as a Magnox reactor due to

the use of a magnesium-non oxidising alloy used as fuel casing. A

schematic of the plant design is shown in Figure 1.1. The reactor used

Pile Grade A (PGA) graphite as a moderator. This moderator was

designed to take advantage of the structural properties of graphite and

Page 24: Using Intercalation to Simulate Irradiation Damage of

24

to provide support for fuel pins, control rods and channels for a carbon

dioxide coolant[3].

Figure 1.1 Magnox Schematic[3]

1.1.3. Advanced Gas Cooled Reactor

The 2nd generation of commercial reactors commissioned and built in

the UK, was the Advanced Gas Cooled (AGR) reactors, Figure 1.2. These

reactors had a number of improved design features. Designed to

operate at ~ 650°C to improve the thermal efficiency, the AGR’s also

employed a re-entrant flow’ system which kept the graphite core

temperature below 450oC in order to avoid any thermal oxidation of

the graphite in CO2 coolant. The reactor used a semi-isotropic graphite

to improve the continued structural integrity of the associated

components. The reactor employed improved fuel and safety features.

The combined result is a significantly improved reactor design with an

output of ~ 1MWe [4].

Page 25: Using Intercalation to Simulate Irradiation Damage of

25

Figure 1.2 AGR Schematic [4]

1.1.4. Very High Temperature Reactors

The future of gas cooled graphite moderated reactors presently lies

with the Very High Temperature Reactor (VHTR) concept[5]. Research is

currently being carried out by a number of nations under the Gen IV

consortium to realise the design. The Gen IV consortium is an

international collaborative effort to design the next generation of

commercial reactors. The VHTR is touted to be the closet Gen IV design

to realisation as it builds upon technology implemented in Germany’s

pebble bed High Temperature Reactor (HTR) and the UK’s prismatic

core HTR (DRAGON) designs from the 1960’s[6]. The VHTR concept

design is shown schematically in Figure 1.3 and aims to operate at

temperatures up to 1000°C circulating helium coolant through a heat

exchanger where a secondary circuit provides heat for electricity

generation and process heat. The coolant choice eliminates oxidation as

a significant damage mechanism, only coming into effect under

emergency conditions. The high outlet temperature is expected to give

a thermal efficiency of over 50%. The main research required is for high

Page 26: Using Intercalation to Simulate Irradiation Damage of

26

temperature materials. This design uses a significant amount of

graphite in the fuel element, reactor core and reflectors[5].

Figure 1.3 Very High Temperature Reactor [7]

1.2. An Introduction to Nuclear Graphite

This section covers a lot of basic ground in order that the reader may

better understand the reason graphite is used in a reactor, its structure

and the reason it has this structure. Starting with the physics of

graphite on the atomic scale we see why graphite is used. Looking at

graphite at increasingly larger scales will give the reader an

understanding of the structure of graphite and a brief description of

manufacturing methods will describe the reason for certain structural

features. The structural features have a significant effect on the

material properties and the way they change under irradiation as shall

be described in section 1.3.

Page 27: Using Intercalation to Simulate Irradiation Damage of

27

1.2.1. Cross Section

The primary reason for the presence of graphite in a reactor is to act as

a moderator. A moderator is reactor component which is used to

absorb the high energy of the neutrons emitted from a fissioning

nucleus.

Two properties known as the scatter cross section and absorption cross

section describe the probability of certain outcomes when a particle of

a given energy travels though a volume of material. The scatter cross

section describes the probability that a neutron will collide with an

atom and the absorption cross sections describe the probability of a

neutron being absorbed.

Thermal reactors require neutrons with thermal energies. A neutron

with thermal energy is defined as a neutron with an energy less than

1eV, neutrons are released from a fission event with an energy from a

few KeV to 10 MeV [8]. This is achieved by placing a moderator around

the fuel pins which reduces the energy of neutrons.

Moderators require a low absorption cross section and high scatter

cross section. There are three main choice of moderator material

typically used in nuclear technology; graphite, light water and heavy

water, The cross sections of these are shown in Table 1.1[9].

The absorption cross section of water is the highest and as such can

only be used with enriched uranium. Heavy water is the best moderator

in this respect having the lowest absorption cross section. Heavy water

is very expensive unlike graphite which is relatively cheap. Furthermore

of all the materials mentioned graphite can be heated to the highest

Page 28: Using Intercalation to Simulate Irradiation Damage of

28

temperature before a phase change occurs, phase changes in

moderators can be catastrophic causing incidents such as Chernobyl.

Table 1.1 Moderator Cross Sections

Material Density (gcm-3) σs (b) σa (b)

H2O 1.0 49.2 0.66

D2O 1.1 10.6 0.001

Graphite 1.6 4.7 0.0045

All three materials are used in various reactor designs. Graphite has

other advantages; it is used for structural components, is chemically

inert and remains stable at high temperatures. The absorption cross

section of moderators are affected by impurities within the

material[10].

1.2.2. Structure and manufacture

Observations of graphite under increasing magnifications show that

each level is made of increasingly small constituent parts as shown in

Figure 1.4. The figure shows that the bulk material is made up of an

agglomeration of coke particles in a binder matrix. The binder matrix

and coke particles are made up of ordered and disordered graphite

crystallites respectively. The crystallites are made up of crystals of

graphite, large regions of carbon atoms in a lamellar arrangement. This

section will detail graphite at each different level.

Page 29: Using Intercalation to Simulate Irradiation Damage of

29

Figure 1.4 Graphite Structure[11]

1.2.2.1. Atomic Structure

The hexagonal structure of carbon atoms are held together by sigma

bonds along the basal plane as shown in Figure 1.5. The sigma bonds

are strong (524 kJ/mol) and short (0.141nm) bonding three valance

electrons. The basal planes are held together by a much weaker

interaction arising from the remaining valance electron forming pi

bonds. The force is relatively weak, 7kJ/mol, and therefore long

(0.335nm). Sigma bonds are significantly stronger than pi bonds

because there is significantly more overlap of the electron orbitals. The

anisotropic nature of the bond structure gives rise to high anisotropy of

the mechanical and thermal properties, whilst dimensionally graphite

crystals are many magnitudes larger in length parallel to the basal plane

than perpendicular to the planes.

The a and c axis are shown in Figure 1.5. It is important to note that the

a axis runs parallel to the strong sigma bonds and the c axis runs

parallel to the weaker dipole interactions. The work often references

these axes.

Page 30: Using Intercalation to Simulate Irradiation Damage of

30

Figure 1.5 Atomic structure of graphite crystal[12]

Graphite is predominantly arranged in an ABAB stacking arrangement

termed hexagonal graphite[12]. This is the thermodynamically stable

form of graphite and has a density of 2.25g/cm3.

It is rare to find large regions of perfect graphene planes organised in

an ideal lamellar structure as there are normally a number of defects

present in the graphite crystal structure. The smallest defects are point

defects, such as out of plane atoms, vacancies within the lattice

structure or a combination of the two. These can be shown to exist by

computational analysis[1] and by observing reactivity changes near

defects with electron microscopy[13]. Larger defects can include

multiple interstitials, multiple vacancies or dislocations, normally edge

dislocations as shown in Figure 1.16. Occasionally screw dislocations as

shown in Figure 1.7 are also present.

Defects have also been looked at in detail with a Scanning Tunnel

Microscope (STM) a method which has also shown the presence of

c

a

a

a

c

Page 31: Using Intercalation to Simulate Irradiation Damage of

31

ribbons on the surface of graphite[14]. STM is an experimental

technique in which a voltage is applied between the tip and sample

surface in close proximity allowing electrons to cross the gap. The

probe tip moves across the sample surface and the change in surface

height vaires the current which can cross the gap allowing very high

resolution surface maps to be obtained[15].

Figure 1.6 Edge Dislocation [16]

Figure 1.7 Screw Dislocation[17]

1.2.2.2. Crystal Structure

The nanostructure of graphite crystallites can be seen using

Transmission Electron Microscopy (TEM), Figure 1.8. The high

resolution images obtained show basal plane stacking in graphite

crystals. The presence of nanocracks known as Mrozowski cracks which

run parallel to the basal planes can also be seen [18, 19]. These cracks

form during cooling of the graphitised block.

During manufacture as the material is cooled from the graphitisation

temperature, at a temperature around 1800°C the bulk structure

hardens and the anisotropy of the Co-efficient of Thermal Expansion

(CTE) forms long cracks to relieve stresses parallel to layer planes[20].

The difference in the thermal contraction in the a and c axis generates

Page 32: Using Intercalation to Simulate Irradiation Damage of

32

stresses and below 1800°C there is no significant thermal creep to

relieve the stresses resulting in crack formation[21].

Figure 1.8 TEM of Highly Orientated Pyrolytic Graphite (Image courtesy of Keyen Wen)

1.2.2.3. Polycrystalline Structure

The microstructure of a polycrystalline graphite can be characterised

using polarised optical microscopy as shown in Figure 1.9. The material

is polycrystalline, made up of binder, filler and flour components.

The degree of filler particle alignment depends on the forming process

used during manufacture. Filler particles are made up of well aligned

crystallites with a high degree of crystallinity. Filler particles contain

lenticular microcracks, these are calcination cracks formed during the

calcining process. The binder matrix is characterised by the disordered

nature of the crystallites, Figure 1.10Error! Reference source not found..

Binder is composed of flour, regions of graphitised coal tar as well as

complex ungraphitised regions. Although the flour is made of crushed

filler particles, these do not necessarily align according to the forming

process.

Graphite

crystal

Morosowski

crack

Page 33: Using Intercalation to Simulate Irradiation Damage of

33

Figure 1.9 Gilsocarbon Microstructure (Image Abbie

Jones)

Figure 1.10 Binder Matrix Crystallite Structure (Image

Abbie Jones)

Within the binder matrix a pore structure, known as gas pores develop

during the baking process. Calcination cracks and Mrozowski cracks are

isolated while gas pores may or may not be[22]. Gas pores therefore

provide accesses to the internal microstructure of graphite to gases

which may cause damage under irradiating conditions depending on

their chemical makeup.

Page 34: Using Intercalation to Simulate Irradiation Damage of

34

1.2.3. Manufacture

Graphite for nuclear applications is manufactured to a purity up to

99.999% using the process detailed in Figure 1.11 [20].

Figure 1.11 Graphite manufacturing process [20]

Most commercially produced nuclear graphites use a petroleum coke

filler as this is an easily graphitised material. The filler particle size has

an effect on the bulk structure properties influencing the, porosity,

strength, and crack resistance[23]. Filler particles are made of large

aligned regions of crystals which have a high strength along the axis of

crystal orientation. However, increased particle size tends to increase

porosity in the microstructure and therefore decrease crack resistance

and affect the strength of the bulk material. The coke is heated to

Page 35: Using Intercalation to Simulate Irradiation Damage of

35

remove volatile hydrocarbons. This produces lenticular cracks in the

filler particles.

The coke is mixed with a coal tar binder and formed. A number of

methods can be used to form the graphite including, isostatic moulding

and vibration moulding. In general extruded graphite results in good

strength and brittleness but poor isotropy, isostatic pressing gives

excellent strength and isotropy with poor brittleness whilst

vibromoulding gives good strength, isotropy and brittleness

characteristics[23].

The formed graphite, the green article, is then baked to drive off

further volatile products. The baking process produces gas evolution

pores throughout the binder material.

Finally the graphite is graphitised, heated to ~3000°C, to improve the

crystallinity throughout the entire block[24]. It is also possible to

improve the purity of graphite further through thermal or chemical

means at this stage.

Page 36: Using Intercalation to Simulate Irradiation Damage of

36

1.3. Irradiation Damage

The principle reason for the using graphite in a nuclear reactor is to

slow down fast neutrons. The mechanism by which graphite slows

down neutrons is also the primary cause of damage to graphite

moderators, although thermal and radiolytic oxidation is also a concern

in some reactor designs. Irradiation damage causes changes to the

dimensional, mechanical and thermal properties of the crystals and

bulk structure.

1.3.1. Atomic Scale Damage

During reactor operation fast neutrons are released from the fuel with

a mean energy of 2MeV[25] and follow a path through the graphite

structure colliding with a large number of carbon atoms in transit. The

high energy neutrons pass on a fraction of energy to the incident

carbon atoms (held in place by a binding energy of ~5eV). The excited

carbon atoms behave like high energy projectiles displacing further

carbon atoms causing a cascade effect as shown in Figure 1.12. The

displaced atoms leave immobile vacancies.

Figure 1.12 Displacement Cascade

Neutron

Primary Knock

on Atom Secondary

Knock on Atom

~100Å

~1000eV

500eV

10Å

Page 37: Using Intercalation to Simulate Irradiation Damage of

37

The displaced atoms are mobile and most fill vacancies [25]. Vacancies

are less mobile than single atoms as more energy is required to

rearrange the atoms for a hole to move. The remaining displaced atoms

may fill interstitial positions creating local defects that have a huge

effect on the bulk properties of the graphite. Displaced atoms tend to

coalesce forming lower energy arrangements. The coalesced atoms

force apart graphene layer planes causing an expansion in the crystal c-

axis shown schematically in Figure 1.13 [26], conversely, vacancies

result in a contraction in the a-axis.

Figure 1.13 Growth of interstitial loops[25]

Figure 1.14 Frenkel Pair[1]

There are a number of possible point defects possible, one of the most

common is the Frenkel pair, Figure 1.14. These point defects create

strong cross layer binding which have a significant effect on material

properties. By creating interlayer valance bonds dislocation glide can be

restricted. These pinning points are thought to be removed by high

energy projectiles, the breakup of these bonds rearranges the atomic

structure and releases energy[27].

Singular interstitial atoms are particularly mobile and can migrate to

form larger less mobile defects. Eventually, a large enough number of

interstitials coalesce such that they may be observed using TEM, Figure

1.15, and can be considered to form a new plane i.e. a prismatic

Page 38: Using Intercalation to Simulate Irradiation Damage of

38

dislocation[28]. Observations using STM have found these clusters to

vary in shape from circular to linear[29].

Figure 1.15 Interstitial loops created by irradiation at 1350°C to 11.7 x 1020 neutrons cm-2 [30]

Each displaced atom leaves a corresponding vacancy. Vacancies are

considered to be mobile, however the barrier energy to such mobility is

significantly higher than for interstitial atoms because more atoms

must rearrange their position for a vacancy to move than are required

for interstitial migration. Therefore a large number of single vacancies

maybe formed. As dose increases multi-vacancies are formed[30].

Multi-vacancy interstitials may reduce or eliminate dangling bonds and

so reduce the energy of such arrangements. Multi-vacancies are closed

in two possible ways; Circular vacancies are filled by the contraction of

planes above and below the vacancy site. Linear vacancies contract in

plane closing along the line of vacancies causing shrinkage in the a-axis.

Over a core lifetime there is a significant amount of annealing, so much

so that it is likely that each atom is displaced at least once[1]. The

Page 39: Using Intercalation to Simulate Irradiation Damage of

39

overall effects of these changes to the atomic structure cause growth in

the c-axis and shrinkage in the a-axis[31].

Recently it has been suggested that there are problems with the

standard model just described. The authors cite a lack of high

resolution microscopy evidence and discrepancies between the

theoretical and measured values of dimensional change and interstitial

migration energy[32]. Instead two mechanism are proposed, at low

temperatures buckles in the graphite plane are pinned by point defects.

At high temperatures pinning is less prevalent allowing dislocations to

interact more freely, when dislocations of opposite sign meet there is

an accumulation of matter which is accommodated by folding of the

layer plane. There is much debate around this theory, even the authors

concede that there is much experimental work required to support the

theory but that it does answer some phenomena the standard model

cannot explain.

1.3.2. Dimensional Change

Highly Orientated Pyrolytic Graphite (HOPG) crystals have been used to

investigate dimensional changes to graphite crystals from fast neutron

irradiation [33]. There is a significantly larger strain induced in the c-axis

than the a-axis as shown in Figure 1.16 and Figure 1.17. At irradiation

temperatures below 300°C there is a volume change associated with

stored energy[34]. The volume change is due to annealing of defects

which release stored energy. Increasing the irradiation temperature

decreases the dimensional change rate because of the increased

thermal annealing of the induced defects. Increasing the final heat

Page 40: Using Intercalation to Simulate Irradiation Damage of

40

treatment temperature and therefore the crystallinity of the graphite,

has been shown to decrease the rate of dimensional change [35, 36].

Figure 1.16 Effect of dose and irradiation temperatures on dimensional change of HOPG in the c-axis[28]

Figure 1.17 Effect of dose and irradiation temperature on the dimensional change of HOPG in the a-axis.[28]

Extruded polycrystalline graphite derived from needle shaped cokes

irradiated at low temperature and dose tend to expand in all directions

Page 41: Using Intercalation to Simulate Irradiation Damage of

41

with greater expansion parallel to the direction of extrusion. Increasing

the radiating temperature decreases dimensional change trends for a

given dose. Irradiating between 150°C and 250°C causes graphite to

expand in the c-axis and shrink in the a-axis, expansion in c-axis

decreases and contraction in the a-axis increases with temperature

[37].

Figure 1.18 PGA Dimensional Change at low Temperature[38]

Above 250°C PGA contracts in both axes with greater shrinkage in the c-

axis. Initial shrinkage is then followed by growth, Figure 1.19, Figure

1.20 and Figure 1.21[31, 33, 39]. In the case of anisotropic graphites

there is a marked difference in the dimensional change of the

perpendicular and parallel directions. The dimensions of isotropic

graphites expand at a rate similar to the perpendicular extrusion

direction of anisotropic graphites.

Page 42: Using Intercalation to Simulate Irradiation Damage of

42

Figure 1.19 PGA Dimensional Changes of PGA at High Temperatures Perpendicular to Extrusion[39]

Figure 1.20 Dimensional Changes at High Temperatures Parallel to Extrusion[39]

Page 43: Using Intercalation to Simulate Irradiation Damage of

43

Figure 1.21 Dimensional changes of Gilsocarbon due to irradiation damage[35]

Above 250°C, the initial contraction of Gilsocarbon is due to contraction

in the a-axis with the filler and flour. Despite the concurrent crystal

expansion in the c-axis, which is of a larger magnitude than the

shrinkage a-axis, the expansion is absorbed by Mrozowski cracks

resulting in an overall shrinkage of the bulk material [40]. Eventually the

Mrozowski cracks close and the continued expansion in the c-axis

becomes the dominant factor resulting in bulk expansion. The point at

which expansion becomes more dominant than shrinkage is known as

turnaround [41]. As the irradiation temperature increases the thermal

expansion of the crystals is larger causing the Mrozowski cracks to close

at a lower dose and by implication turnaround occurs at a lower dose

[42]. Under loading conditions dimensional change is affected by

irradiation creep[43].

Page 44: Using Intercalation to Simulate Irradiation Damage of

44

1.3.2.1. Modelling Dimensional Change

A number of attempts have been made to model and predict

dimensional changes of graphites without the expensive and time

consuming process of using materials test reactors.

Simmons observed that a linear relationship exists between the CTE,

D.C.CTExx, and the dimensional changes due to irradiation, D.C.IRRxx, at

low doses [44].

axcxxxIrr gAgACD )1(..

1. Equation 1.1

axcxxxCTE AACD )1(.. Equation 1.2

where gc and ga denote growth due to irradiation, βa and βc are the CTE

of the crystal in the a and c axis respectively and Ax is the structure

factor relating the two equations. Ax is a function of crystal temperature

dependant CTE and describes contribution of porosity and crystal

orientation to the material expansion. The model only stays true at low

irradiation doses because it assumes that the crystallites are a loose

arrangement with no interaction between them. This model thefore

does not predict turnaround[45].

Sutton and Howard[46] investigated the role of thermal expansion of

bulk nuclear graphite grades further and considered the following three

factors to be important;

a) The thermal expansion of the highly anisotropic graphite crystal.

b) The orientation of the crystals

Page 45: Using Intercalation to Simulate Irradiation Damage of

45

c) Accommodation porosity in the form of Mrozowski cracks and fine

pores generated between adjoining crystallites.

The equations derived from these assumptions describe the coefficient

of thermal expansion experienced by bulk graphite for a given dose and

temperature;

acGW KKCD 21.... Equation 1.3

acGA KKCD 43.... Equation 1.4

where Kx are factors describing the orientation of crystals, σa and σc

describe the single crystal coefficients of thermal expansion while α and

γ are accommodation factors[46]. D.C.W.G. and D.C.A.G. describe the

thermal expansion in the With Grain (WG) and Against Grain (AG)

direction. As shown in Figure 1. 22 WG and AG describe the

predominant grain orientation direction which is determined by the

manufacturing process.

Page 46: Using Intercalation to Simulate Irradiation Damage of

46

Figure 1. 22 Schematic describing the with grain and against grain directions for either pressing or extrusion

forming

Focussing on thermal expansion has been but one of many different

techniques used to investigate dimensional change. Brocklehurst[47]

used the emerging field of intercalation to induce dimensional change.

The feasibility of intercalation was assessed by comparing intercalation

rates to the Simmons relationship. The study showed that it was

feasible to use bromination as a technique to investigate structural

changes[48] and he deduced that microstructural changes such as

crystal expansion and pore generation determined the change in bulk

properties.

The finite element method was used by Hall[49, 50] to create a multi-

scale model of graphite. By inputting irradiation data of HOPG into a

crystallite model and using the output as input parameters for a bulk

material model. This method was quite successful in predicting changes

seen in reactor conditions. It was concluded that to verify his model it

Page 47: Using Intercalation to Simulate Irradiation Damage of

47

must be determined if crack closure does occur under reactor

conditions [51].

1.3.3. Young’s Modulus

Generally in engineering Young’s modulus is measured by performing

static loading experiments and observing the stress strain curve. The

stress strain curve of graphite is non-linear and this poses problems for

measuring Young’s modulus. Figure 1.23 shows that very quickly there

is plastic like deformation with hysteresis in the stress strain curve. For

nonlinear materials it is possible to define the modulus by a chord

length[52], unfortunately there has been little consistency in doing this

by graphite researchers. It is most common to measure the Young’s

modulus dynamically. Dynamic Young’s modulus (DYM) is a method of

inducing strain using ultrasound and is approximately the rate of

change at the origin[53].

DYM is measured by sending an ultrasonic elastic wave through a

material which generates small strains as it travels. Therefore static

tests only produce similar results if carried out for very small elastic

deflections. Ultrasonic elastic waves are a very versatile technique for

materials testing and further to the measuring the modulus of a

material can also be used to gain information of defects within the

material by measuring parameters such as signal attenuation and the

time for echo return.

Page 48: Using Intercalation to Simulate Irradiation Damage of

48

Figure 1.23 Nuclear graphite stress strain curve[54]

The elastic constants of graphite are often referred to using Voight

notation as given in Table 1.2[55] and shown schematically in Figure

1.24. Figure 1.24 shows the orientation of the unit vectors in relation

to graphite layer planes. C11, C22 and C33 describe Young’s modulus in

the a, a, and c directions respectively whilst C44 describes the shear

modulus perpendicular to the layer planes and C66 describes the shear

modulus parallel to shear planes.

Table 1.2 Voight notation

2. Unit

Vector

11 22 33 23=32 31=13 12=21

Voight

Abbreviation

1 2 3 4 5 6

Page 49: Using Intercalation to Simulate Irradiation Damage of

49

Young's modulus of graphite crystals can be calculated from elastic and

crystal compliances. The values of virgin graphite in the c-axis are

around 1024 GPa in the a-axis and 36.4 GPa in the c-axis[51]. A small

initial irradiation dose greatly increases single crystal modulus, Figure

1.25. Simmons hypothesised that this is due to pinning of the basal

planes causing an increase in the shear modulus C44. Further exposure

leaves a plateau in the modulus and it has been assumed that this is

due to a saturation of the number of pinning points created[43] [1].

Saturation is thought to occur because after a certain dose a large

number of holes exist within the structure making it likely that a

displaced atom will recombine with a vacancy. Values of the elastic

compliances S11, S12 and S13 are essentially unchanged during irradiation

[56].

31 13

21

33

32

11 23

22

12

Figure 1. 24 Orientation of unit vectors in relation to graphite layer

planes

Page 50: Using Intercalation to Simulate Irradiation Damage of

50

Figure 1.25 Change in crystal shear modulus C44 due to irradiation at 50oC[56].

Note 1dynes/cm2 is equal to 0.1 Pa

Modulus measurements of polycrystalline graphites also exhibit the

initial increase due to the change in crystal shear modulus, Figure 1.26.

It is thought that the increase in crystal shear modulus increases the

bulk modulus of graphite, this is due to experiments which show only

C44 shows any significant change due to dose[56]. The explanation given

is that the reduction of basal plane slippage implied by a high C44

prevents graphite from deforming so readily.

At higher dose a secondary increase is observed and this is considered

to be due to closure of Mrozowski cracks[57]. The closure of cracks is

thought to increase the modulus as the decrease in internal porosity

increases the stress required to induce a given compression strain, as it

is easier to compress a porous volume than a solid volume.

After this the modulus decreases rapidly as the microstructure

disintegrates. Increasing irradiation temperature decreases the

magnitude of the secondary increase and the dose required to reach

the peak of the secondary increase. Low dose Changes in Young's

modulus can be annealed out below 1000°C corresponding to

Page 51: Using Intercalation to Simulate Irradiation Damage of

51

aggregation and annealing of small interstitial loops [58]. Changes in

modulus are shown to be significantly affected by irradiation creep [59].

Figure 1.26 Change in Gilsocarbon Young's modulus due to irradiation[35, 39]

There have been fewer attempts to model the change in modulus

though Hall[49, 50] had success with multi-scale finite element analysis.

Using the code ABAQUS a model of a graphite crystal was created and

the effect of irradiation on the crystal properties due to irradiation such

as dimensional change and co-efficient of thermal expansion were

input. The output of this model was input into a polycrystalline model.

The results showed that an increase in young’s modulus could be

attributed to closure of microporosity and that at higher doses inducing

microcracking reduced the modulus as seen in experimental data .

1.4. Scientific Proposal

This project will use intercalation of bromine as a means of inducing

structural damage to simulate irradiation damage in nuclear grade

graphites. In particular this work will focus on the anisotropic graphite

Page 52: Using Intercalation to Simulate Irradiation Damage of

52

PGA and semi isotropic graphite Gilsocarbon. Work shall also be carried

using HOPG to model single crystal intercalation. Investigating both

polycrystalline and single crystal graphites will provide an insight into

the interrelationship of properties in these materials. Using

intercalation as a simulation technique has the advantage over

irradiation for PhD investigations as large strains can be induced

relatively quickly, over the course of a few days rather than a few years.

This technique also means standard laboratory equipment can be used

as there is no need to accommodate radioactivity.

1.5. Summary

Nuclear graphite has been used since the very first nuclear reactors as a

moderator. This is because it has a high scatter and low absorption

cross section. When graphite is placed in an irradiating environment

severe structural damage occurs. This causes the properties of graphite

to change. The manner in which they change is affected by the

microstructure of graphite. Nuclear graphite grades have different

microstructures and this is due to the manufacturing process. This work

will use intercalation to investigate selected material property changes

in nuclear graphite.

Page 53: Using Intercalation to Simulate Irradiation Damage of

53

2. Intercalation

The core of a nuclear reactor is an extremely hostile environment and

this poses significant problems for designers and operators. High

neutron fluxes, high temperatures and corrosive coolant gases damage

the very materials from which a reactor is constructed. A large portion

of a reactor is constructed from graphite bricks, these bricks receive

substantial atomic scale damage which induces dimensional and

material property changes to both single crystal and bulk polycrystalline

graphite. As discussed in Chapter 1 it is considered that a very

significant portion of these changes are due to the growth of interstitial

layers generated by atoms displaced during collisions[60-62].

Intercalation is another method of inducing damage into the graphite

microstructure and though there are key differences in the damage

mechanisms intercalation of bromine into graphite has been used in

previous studies to simulate irradiation damage[63]. This chapter

discusses the previous work undertaken and is aimed at understanding

intercalation compounds, the chemical process and in particular,

graphite intercalated bromine. The chapter also focuses on methods of

inducing intercalation, the structure of intercalation compounds, the

intercalation reaction and the material properties of the intercalation

compounds.

2.1. Intercalation Compounds

Intercalation has been utilized by humans for many centuries as the

process by which clay can be transformed from a plastic medium to a

Page 54: Using Intercalation to Simulate Irradiation Damage of

54

dense and brittle material for pottery[64] with some of the oldest

known ceramic objects dated as 26000 years old[65].

The first scientific investigation into intercalation was carried out in

1841 when Schauffautel intercalated Sulphate ions in graphite[66]. The

advent of X-Ray Diffraction (XRD) made it possible to carry out more

detailed investigations into these novel materials[67]. Around the same

time the science progressed further in the development of new

intercalation compounds[68].

The definition of intercalation is insertion and literally refers to the

insertion of time into a calendar. In chemistry it refers to the reversible

insertion of a guest compound into a host structure[66] where the

intercalate fills voids in the atomic structure. There are many possible

host materials including isotropic lattices such as Zeolites[69],

Orthotropic lattices such as graphite[70] and one dimensional

structures such as DNA[66].

The intercalation of many different allotropes of carbon have been

studied including graphene[71], graphite[66, 70], diamond[72],

fullerenes[73] and carbon nanotubes[74]. Intercalated materials have

been shown to have remarkable properties including

superconductivity[75]. These materials have received such interest over

recent years as there are many realized and potential technical

applications including catalysts, gas sensors, batteries,

superconductors, micro-mechanical actuators and lubricants.

Page 55: Using Intercalation to Simulate Irradiation Damage of

55

2.2. Graphite Intercalation Compounds

Graphite is unique among host materials in that it may be an electron

acceptor or an electron donor[76] and intercalants are classified as

such. Donor compounds include alkali metals, lanthanides and metal

alloys. Ternary donor compounds have also been prepared, normally

alkali metals bonded with molecules such as ammonia[77] or

benzene[78]. Many acceptor compounds have been prepared and are

often based on Lewis acids, that is a species that accept lone pair

electrons, such as bromine[79]. Metal halides (including metal

chlorides, bromides, fluorides) acidic oxides and acids such as sulphuric

acid may also be intercalated and likewise fall into the acceptor

category[70]. The majority of these compounds are unstable in air,

donors are easily oxidized and acceptors readily desorb. Graphite-FeCl3

and SbCl5 are relatively stable and are therefore used for many

experiments which aim to understand intercalation. The stability of

samples can be improved by cooling to liquid nitrogen

temperatures[70].

2.2.1. Intercalation Methods

Graphite can be intercalated with solids, liquids and gases [80];

however it is preferable to intercalate with the gaseous form of the

intercalant if possible as this allows the highest degree of control over

the experimental conditions and the technique is the simplest to

perform. Figure 2.1 shows an early experimental arrangement as used

by Brocklehurst[63] for investigations into intercalation of nuclear

graphite. A graphite sample is held on a calibrated spring in an

evacuated chamber. The gaseous intercalant is released by breaking

Page 56: Using Intercalation to Simulate Irradiation Damage of

56

vials containing the appropriate gas in an adjacent chamber connected

by a stopcock. On opening the stopcock the intercalant quickly fills the

sample chamber. The stopcock is closed and a fresh vial can be added

to the intercalant chamber. As more vials are added the partial pressure

of bromine increases[63].

Figure 2.1 Bromine intercalation apparatus for weight uptake measurements[63]

There are variations on how the intercalant partial pressure is

controlled. Figure 2.2 details the experimental setup Sasa used to

investigate the staging of Graphite Intercalation Compounds (GIC’s).

Page 57: Using Intercalation to Simulate Irradiation Damage of

57

Here the intercalant chamber is placed in a temperature controlled

water bath. Increasing the temperature of the bath increased the

partial pressure of the gaseous intercalant [81]. This method gives

much closer control of the intercalant partial pressure.

Figure 2.2 Bromine Intercalation with control of intercalant partial pressure for XRD measurements[81] (1) Glass

sample holder; (2) a hole of 1mm diameter; (3) graphite sample; (4) “Teflon” sheet (0.2mm thickness); (5)

ground-glass contact: (6) Araldite sealing; (7) bromine (8) water bath for bromine pressure control; (9) to be

joined to vacuum system

2.2.2. Structure of Graphite Intercalation Compounds

The structure of graphite intercalated with bromine was first

investigated by Rudorff [79], who found that bromine intercalated up

to a maximum ratio of C8Br. Importantly the analysis carried out by XRD

showed that galleries of bromine were created and interestingly these

galleries were very ordered. In the case of C8Br there are two graphite

Page 58: Using Intercalation to Simulate Irradiation Damage of

58

layers followed by a layer of bromine. This is known as a second stage

intercalate as shown in Figure 2.3. It has since been shown that

graphite has further lower stages, above C16Br the third stage is created

[82] and a fourth stage exists above C28Br [83]. Different intercalation

species can be intercalated to higher and lower stages, Figure 2.4 shows

a micrograph of a stage 2 CuCl2 graphite intercalation compound. The

CuCl2 is indicated by the white lines and the graphite is indicated by the

dark lines. The image shows that ordered staging can exist over a large

range.

Figure 2.3 Schematic of increasing intercalation stages

Figure 2.4 TEM Image of CuCl2 (light regions)

intercalated graphite (dark regions) [84]

When intercalated, adjacent graphite layers retain their stacking

sequence[70]. Depending on the size of the intercalation species the

induced change in layer plane spacing differs. Brocklehurst compiled

data of a number of different intercalates from a number of different

sources, and is shown in Figure 2.5. This provided Brocklehurst with his

primary reason for using bromine as the intercalation compound for his

experiments which is that the inter-carbon layer distance is

Stage 4 Stage 3 Stage 4

Page 59: Using Intercalation to Simulate Irradiation Damage of

59

approximately the same, 7.05Å, as if the separation was caused by

interstitial carbon atoms, 6.70Å.

Plot of Graphite C-C Interlayer Apacing Containg A Layer of

Intercalated Reactant against Ionic Radii of Reactant

4

5

6

7

8

0 1 2

Ionic Radius (A)

C-C

In

terl

ay

er

Sp

ac

ing

(A

)

Na

K

F

RbCs

Br

Figure 2.5 Distance between graphite layer planes for different intercalation species [63]

Intercalates further modify the structure of graphite by inducing basal

plane slippage. Stage two graphite has AB|BC|CA|AB stacking whilst

stage three compounds have ABA|ABA|ABA stacking where the |

denotes an intercalant layer as shown in Figure 2.6 [85]. This is shown

schematically in Figure 2.6. A similar basal plane slippage is induced by

interstitial carbon atoms[1].

a)

b)

Figure 2.6 Graphite basal planes (black) with intercalant (white) a) stage 3 b) stage 2

The arrangement of bromine at low pressure has been studied using

density functional theory (DFT) Figure 2.7. The molecular arrangements

detailed in (a), (b) and (c) show bromine molecules normal to the

graphite plane, these arrangements are only possible on the outermost

Page 60: Using Intercalation to Simulate Irradiation Damage of

60

graphene plane. Arrangements (d) to (h) show possible bromine resting

positions in between graphene planes. The number indicates the

relative stability of each arrangement in eV (a) is actually 0.004eV [71].

The unit cell of bromine intercalated graphite has been widely studied

using electron diffraction [86], x-ray diffraction [87] and neutron

diffraction [88], [89] Figure 2.8. The in plane unit cell is base-centred

orthorhombic. When bromine is intercalated it maintains its molecular

identity, that is it maintains the bromine – bromine bond length seen in

solid bromine [90].

Figure 2.7 Bromine positions at low concentrations calculated

by DFT [71]

Figure 2.8 Unit Cell C8Br found by Neutron

Diffraction [91]

Figure 2.9 and Figure 2.10 present electron micrographs of damage

sites within brominated and irradiated graphite. The size and quantity

of irradiation clusters are strongly related to irradiation temperature.

As irradiation temperature increases irradiation clusters become more

widely separated and larger [92]. By comparison TEM has shown that

the damage sites of brominated graphite are larger than irradiated

graphite and that there are fewer damage sites per unit area. The

distribution of intercalation damage sites appears to be related to the

Page 61: Using Intercalation to Simulate Irradiation Damage of

61

surrounding microstructure, local faults and defects [27], whereas

irradiation defects are more randomly distributed[30].

Figure 2.9 Intercalate clusters (bromine main picture

Iodine chloride inset)[27]

Figure 2.10 Interstitial clusters caused by fast

irradiation[30]

2.2.3. The Intercalation Reaction

Understanding the intercalation reaction is fundamental to

understanding intercalation as a technique to simulate irradiation. This

is particularly true when using it for microstructural investigations as

this thesis aims to do. It is important to understand where the reactant

is most likely to be located and how the nature of the host and reactant

may affect this.

The amount of reactant intercalated is related to the partial pressure of

the reactant. The most detailed study was carried out by Sasa who

performed an intricate experiment to investigate the effect of partial

pressure on the reactant/host ratio. Sasa measured the amount of

bromine intercalated for a given partial pressure and in conjunction

with xray diffraction measurements deduced the pressures required for

each phase change from pure pyrolytic graphite to a fully intercalated

compound, this is detailed in Figure 2.11.

Page 62: Using Intercalation to Simulate Irradiation Damage of

62

The X-ray diffraction results show that as the partial pressure increases

the weight uptake increases and associated with the weight uptake is a

change in phase. A change in phase is followed by an increase in the

rate of intercalation, this is due to the intercalant arranging in

increasingly dense arrangements. On removal of the intercalant there is

a hysteresis associated with intercalant removal. On deintercalation it is

not possible to remove all of the intercalant unless the host is heated to

suitably high temperatures[20, 93].

Figure 2.11 Phase change induced by change in intercalant pressure. Open circles indicate intercalation and

closed circles denote deintercalation[81]

Ubbelohde et al[94] examined the bromination initiation process and

proposed that the intercalant atom may become adsorbed on the

graphite surface, thus forming an ionic bond with a carbon macro-

molecule. It was thought that this change in charge distribution would

“unpin” the graphite layers and allow intercalation to occur by

diffusion. Recent thermogravity experiments on graphene confirm the

importance of chemisorption [95].

Ubbelohde’s idea suggests that once graphite has adsorbed the

intercalant any layer is equally likely to intercalate. The theory was

Page 63: Using Intercalation to Simulate Irradiation Damage of

63

shown to be incomplete by a very interesting experiment by J.G.

Hooley[96]. By taking 60mm thick piece of pyrolytic graphite with five

equally spaced marks along its height (parallel to the basal planes) he

could make strain measurements at five separate regions of a sample

as intercalation took effect. The results shown in Figure 2.12 show that

the rate of intercalation is quickest at the outermost regions of the

sample to a point of saturation with intercalation of the inner regions

being slightly delayed and that therefore not every layer was

immediately open to intercalation. Later theoretical work showed that

this effect was due to a long range elastic interaction between

intercalant layers [97, 98].

Figure 2.12 Intercalation Rate through sample thickness [96]

An interesting area of research regards the way an intercalation

compound changes from one stage to the next. The problem arises in

the higher stage compounds when the insertion or removal of an

intercalant layer will not result in the regular intercalant spacing seen

with intercalation compounds. There is a change in the XRD spectra

associated with the phase change which shows the rise and fall 00l

peaks. It is very difficult to determine from these how the intercalation

Page 64: Using Intercalation to Simulate Irradiation Damage of

64

compound changes from one phase to another, though there are two

proposed theories.

Nixon et al studied intercalation of nitrates and interpreted their results

as showing whole layers deintercalating and intercalating at the

appropriate level. They suggest this is achieved by mobile dislocations

which are present in all layers at the edges of graphite and travel

towards the centre at the appropriate intercalate concentration. Nixon

et al reasoned that this will produce the most stable arrangement [85].

A second theory has been proposed by Daumas and Herold who found

it unlikely that whole layers of intercalate would be removed before

new layers could intercalate to achieve the next stage. Either stages

had to be missed or there was another way. Daumas proposed that

there are microscopic domains of well staged material; however the

domains may be in different layers of graphite as seen in Figure 2.13. As

the surrounding intercalates partial pressure changes, the domains

move along the layer planes such that the different stages are formed.

Theoretical work shows that a long range elastic interaction between

intercalant islands will drive the random or mixed staged compounds to

pure stage ordering[97, 98].

Figure 2.13 Stage two (left) and stage three (right) Daumas Herold Domains[99]

Page 65: Using Intercalation to Simulate Irradiation Damage of

65

The domain model is supported experimentally with evidence collected

by Clarke et al[100]. By intercalating graphite with potassium to a stage

2 compound and applying a load, XRD measurements showed that two

stages can coexist in a sample and that there is a time dependant

change in the super lattice order. Clarke noticed that the degree of

crystalline perfection has a significant effect on the load under which

two stages can coexist, suggesting this may be from differences in

dislocation densities and the associated difference in shear modulus,

C44. Evidence of mixed stages has since been observed by TEM in

graphite samples intercalated with FeCl3 [101], K[102] and CuCl2 [84].

A TEM study on graphite with residue bromine, SbCl5 and KHg

intercalates by Timp & Dresselhaus [103] found large domains, up to

25nm x 210nm, and using diffraction contrast techniques found these

domains to be free of dislocations. By looking at the resulting

interference between electron waves which have passed through the

sample using a TEM microsope it is possible to observe dislocations in

the atomic structure. Timp and Dresselhaus looked for basal plane

dislocations which would be indicative of the Damaus and Herold

model but could not find any. This runs contrary to the domain model

and Timp & Dresselhaus suggest that while the Daumas & Herold model

may hold for other intercalates it does not for those chosen for this

experiment. They suggest the stage changes involve the intercalate

diffusing through point defects in the graphite lattice allowing the

movement of full intercalate layers. A later study into SbCl5 using a

Scanning Ion Microprobe did find domains of intercalant on freshly

cleaved sample surfaces [104].

Page 66: Using Intercalation to Simulate Irradiation Damage of

66

Axdal and Chung[99] developed a theory that relates results from many

different intercalation experiments of different intercalation

compounds. The reaction is split up into a number of rate determining

steps. The initial step occurs while the intercalant is external to the

graphite, that is evolution of the intercalate and transport of the

intercalate to the sample. Then surface reactions occur, adsorption to

the surface and nucleation of insertion sites. This is followed by

diffusion of the intercalant through the sample. The final step is the

staging reaction. The overall rate of reaction will strongly depend on

which of these steps takes the longest. They show theoretically that in

the case of bromine, diffusion is the rate controlling mechanism. This

implies that the chemical potential for a reaction decreases towards the

centre of the sample. This agrees with experimental work carried out in

parallel by the authors[105].

Intercalation is strongly related to the degree of crystal perfection

affecting both the rate and amount of intercalant uptake. A study by

Ubbelohde et al[106] investigated how the graphitisation temperature

used in the formation of pyrolytic graphite affected the uptake of

bromine after 70 hours as shown in Figure 2.14. The study shows that

above 1900°C there is a significant decrease in the c-spacing coupled

with a decrease in stacking disorder suggesting a step change in the

number of defects present. The reduction in defects allows the graphite

to absorb bromine to its maximum ratio C8Br [106, 107]. This is

important as it indicates the degree of crystal perfection has an effect

on the uptake of intercalate and this may potentially have an effect in

polycrystalline graphite where different regions have graphitised to

different degrees.

Page 67: Using Intercalation to Simulate Irradiation Damage of

67

Hooley[107] showed that the rate of bromine uptake was also directly

related to crystal perfection with the samples graphitised at a higher

temperature intercalating at a faster rate. It has also been shown that

intercalation is not affected by chemical impurities in a graphite lattice

[108].

Figure 2.14 Effect of crystal perfection on bromine uptake [106]

A detailed microscopy study by Heerschap[27] investigated dislocations

induced by bromine intercalation. The boundary dislocations of

intercalated bromine are sessile, that is they can only travel in the glide

plane by diffusion of the intercalant. The boundaries are highly mobile

and have a low line tension. Evidence is presented that bromine easily

travels through line defects present in the original graphite crystal,

widening the dislocation as more bromine is intercalated.

In summary the rate of bromine uptake is affected by the degree of

crystal perfection of the graphite and the partial pressure of the

surrounding intercalant with the reaction unable to initiate if either of

these two components are insufficient. The reaction is initiated by an

Page 68: Using Intercalation to Simulate Irradiation Damage of

68

unpinning of the graphite layers by the intercalant drawing charge. The

reaction is a diffusion controlled process and is progressed by a long

range stress interaction. Staging is a phenomenon unique to

intercalation compounds where long range order is seen and it is a

matter of debate how staging progresses. Experimental work has seen

multiple stages co-exist which strongly suggests domain interactions

are the mechanisms behind stage changes. However it is disputed

whether this is the mechanism in a graphite bromine intercalation

compound.

2.3. Property Changes

2.3.1. Dimensional Changes

Dimensional changes are the most studied property change of

intercalation compounds with investigations into both single crystal and

polycrystalline compounds. A significant portion of the work which

investigated the intercalation mechanism was carried out on single

crystal graphites with dimensional change being a key measurement

[81, 85, 94, 106, 107, 109, 110].

The first studies into the intercalation of nuclear graphite were

concerned with dimensional change. Brocklehurst [63] found that

polycrystalline graphites have an initial linear growth per unit

concentration which could be modelled by Simmon’s growth

theory[111], that is the thermal expansion of graphite can be used to

predict the initial dimensional change behaviour under intercalative

conditions. The relationship held true for single crystal graphites but

not for poorly graphitised materials which were shown to intercalate

Page 69: Using Intercalation to Simulate Irradiation Damage of

69

poorly. After the initial predictable linear expansion, polycrystalline

graphites expand at an increased rate and it is hypothesized that this is

due to opening of porosity by misaligned crystallites. The subsequent

increase in expansion was shown to be less pronounced for isotopic

graphites [63].

An interesting study has been carried out relating the microstructural

changes in battery electrodes to the bulk material dimensional changes

during intercalation of lithium. The graphite electrode material is

polycrystalline made of binder and filler components as with nuclear

graphite, although the associated filler particles and porosity are

significantly smaller, with average pore size around 25µm and 10µm

respectively [112]. This research concluded that the dimensional

changes seen in lithiated electrode graphite are large in the filler

particles and an order of magnitude lower in the bulk material. DFT

calculations show an interstitial lithium atom shares its charge with the

twelve nearest carbon atoms which strengthens interlayer bonding

causing a change in C33 and C44. At the atomic scale an increase in bond

strength is expected to increase the modulus, however at larger scales

this is complicated by the presence of porosity in the bulk structure. It

has been postulated that the change in modulus has a significant effect

causing contraction and expansion of surrounding regions of binder

which can have a different modulus [113].

2.3.2. Young’s Modulus

The data on change in elastic constants of bromine intercalated

graphite is sparse though measurements have been taken by neutron

Page 70: Using Intercalation to Simulate Irradiation Damage of

70

diffraction reporting a decrease in the shear modulus C44 from 0.89 x

109 N/m2 to 0.18 x 109 N/m2 [88].

An interesting DFT study of lithiated graphite [114] has looked at the

effect of lithium intercalation on graphite electrodes, Figure 2.15. It has

been shown that as the stages lower, the linear elastic constant C33

increases by two and a half times and the shear elastic constant C44

increase by up to four times. By applying the results to a polycrystalline

model the Young’s modulus is calculated to increase by a factor of

three and the poisons ratio is calculated to fall from 0.31 to 0.24.

The Poisson’s ratio of nuclear graphite is lower than that of battery

electrodes measured as 0.21 for Gilsocarbon and 0.07 for PGA[35].

Measured changes due to irradiation are noisy[115] and are therefore

assumed to be unchanged by irradiation[42, 115]. As with irradiated

graphite, a change in C44 has also been attributed to the initial rise in

Young’s modulus[56].

Figure 2.15 Change in elastic constants of lithiated graphite[114]

Page 71: Using Intercalation to Simulate Irradiation Damage of

71

Existing data on modulus changes of intercalated materials is sparse,

therefore this literature review was expanded to include intercalated

graphite fibre epoxy composites. Research shows that there is little

change in modulus between the two materials, though authors report

that these property measurements are controlled by the epoxy fibre

interface properties and so provide poor reflection on property changes

due to intercalation[116, 117].

2.4. Summary

Bromine has been used in previous studies to simulate irradiation of

graphite, chosen for its ability to produce similar layer plane spacing to

interstitial carbon atoms. Intercalation refers to the reversible insertion

of a substance in to a host material. There are a range of host materials

which require voids in their structure to accommodate the intercalant,

to which carbon based materials are particularly suitable. It is possible

to intercalate many different substances, with respect to graphite.

These intercalants can be donors or acceptors depending on charge

transfer. When materials are intercalated they form staged structures,

that is the intercalant will fill regular periodic positions perpendicular to

the layer planes.

Bromine uptake is related to the partial pressure of the surrounding

intercalant and the crystal perfection of the host. Intercalation is

initiated by an unpinning of carbon layers and progressed by elastic

interactions. It is unclear if the intercalate fills whole layers or islands

within the host; evidence has been presented for and against each idea.

Axdal and Chung[99, 105] have shown that bromine intercalation in

Page 72: Using Intercalation to Simulate Irradiation Damage of

72

graphite is a diffusion controlled process and that therefore there is less

potential for a reaction towards the centre of a sample.

The most studied property change induced by intercalation is

dimensional change, this has also been modelled extensively on single

crystal graphites and shown to cause an expansion parallel to layer

planes. Various polycrystalline graphite grades have also been studied

by Brocklehurst[63] who showed that initial expansion could be

modelled using Simmons growth theory and he hypothesised that a

later increase in expansion rate could be due to the opening of

microporosity by misaligned crystallites.

Data on young’s modulus of intercalation compounds is sparse, neutron

diffraction investigations of brominated graphite suggest there is a drop

in the shear constant C44 of HOPG. A theoretical study of lithiated

graphite shows a significant increase in elastic constants.

Page 73: Using Intercalation to Simulate Irradiation Damage of

73

3. Materials and Methods

This study is concerned with investigating microstructural damage in

graphite and the effect this has on the bulk material properties. This

chapter details the methods used in this investigation including the

experimental techniques used and design of experimental rigs.

3.1. Experimental Techniques

3.1.1. McBain Spring Balance

As mentioned in Chapter 2 a common piece of experimental equipment

used for intercalation experiments is the sorption balance, otherwise

known as a McBain spring balance [118]. The method is very simple

consisting of a specimen attached to a calibrated spring as shown in

Figure 3.1. A travelling micrometer is used to track the displacement of

the spring, and using Hookes law the change in mass can be calculated,

Equation 3.1.

1hkm Equation 3.1

where m is the mass, k the spring constant in g/mm and h1 is the

position of the bottom of the spring. When using this technique care

must be taken to prevent external interference with the spring.

Vibration of the rig must be prevented and the gas must enter the

chamber at a slow rate[119] to avoid turbulence.

Page 74: Using Intercalation to Simulate Irradiation Damage of

74

Figure 3.1 Schematic of McBain Spring Balance[118]

The McBain spring balance is designed for measuring changes in

sorption with temperature so Brocklehurst [28] modified the McBain

spring balance [118] for studies into the dimensional change associated

with intercalation of graphite. The difference being that there was no

temperature control and the strain induced by sorption was measured.

This was achieved by ensuring the sample dimensions were such that

the travelling micrometer could be used to measure the top and

bottom of the samples to obtain the change in sample length.

3.1.2. X-ray Computed Tomography

X-ray Computed Tomography (XCT) is a non-destructive three

dimensional imaging technique. This study utilises micro tomography to

study the development of intercalation and associated strains within

polycrystalline graphitic microstructures. Strains are measured using

Thermometer

Thermometer

Liquid to be employed

for sorbption

Sorbent

Silica spring

Heating element

Page 75: Using Intercalation to Simulate Irradiation Damage of

75

Digital Image Correlation (DIC). This section is divided into two; the first

part describes the experimental setup of tomography and how to

obtain good quality data; the second part describes the image analysis

techniques used to measure strains induced by intercalation.

3.1.2.1. Experimental setup

The experimental set up for an x-ray tomography experiment is shown

in Figure 3.2. An X-ray source beams photons through a sample. The

sample is rotated through at least 180° with an image taken at suitable

rotational increments. The x-rays are normally passed through a

magnifying optical set up and then converted by a scintillator into a

form which can be detected by a Charged Coupled Device (CCD) camera

[119] and recorded as radiographs. Data may then be reconstructed

and analysed by software.

Figure 3.2 Tomography setup

X-rays can be produced by laboratory sources and synchrotron sources.

Laboratory equipment produces x-rays by firing electrons at a solid

metal anode, radiation is emitted as the electrons are decelerated. This

produces photons with a range of wavelengths [120]. Synchrotrons

produce x-rays by diverting the path of an electron beam through a

magnetic field; this produces a polychromatic beam of x-rays which if

Page 76: Using Intercalation to Simulate Irradiation Damage of

76

desired can then be passed through monochromating optics to produce

a monochromatic beam[121]. The effect of this is that Synchrotrons

have monochromatic light at high flux while laboratory equipment has

polychromatic light at low flux. The higher flux achievable with

synchrotron sources means data can be acquired at a significantly

quicker rate and so can be used for time series investigations.

X-rays are ideal for tomography as they penetrate solid materials and

by measuring x-ray attenuation it is possible to gather information

about the internal structure of a sample. The x-ray absorption cross-

section varies depending upon the atomic structure of the different

phases of the material under investigation and the energy the photons

emitted from the x-ray source. The detector picks up a shadow of a

sample which represents the difference in intensity due to variation in

attenuation across the sample. Figure 3.3 shows the effect of different

sample variations on the x-ray shadow as measured by a radiograph.

Page 77: Using Intercalation to Simulate Irradiation Damage of

77

Figure 3.3 Effect of sample composition on x-ray shadow

When conducting an x-ray experiment a number of difficulties may

arise. The first problem is due to the statistical nature of photons, there

is a probability associated with photon production and interaction with

both sample and detector which causes a speckled effect on the

reconstructed images. There are two options for overcoming this issue,

first method is to increase the length of time for each projection, thus

increasing counting statistics; the second is to use a smoothing

algorithm. A smoothing algorithm will result in a loss of data to a

certain degree however time restrictions normally require a

compromise[122].

Another experimental difficulty may arise from beam hardening which

is a particular problem when using polychromatic light sources. One of

Page 78: Using Intercalation to Simulate Irradiation Damage of

78

the factors which affects x-ray attenuation is the x-ray energy, generally

the lower the x-ray wavelength, the higher the attenuation is. This

results in low energy light to be disproportionately absorbed through

the sample thickness, this manifests as a darkening towards the sample

centre and may be overcome using various algorithms [122].

There are also issues related to the detectors as not all the pixels will

give the same output for a given photon count. This will show up as

light and dark regions in a radiograph. This is compensated for by

calibrating the equipment taking a series of dark and white images

[122].

Nuclear graphite has been studied extensively by tomography. The first

study investigated density variations of thermally oxidised IG110, at

that time a candidate material for Japans High Temperature Gas cooled

Reactor (HTGR)[123]. Further work was carried out at the University of

Manchester on tomography of thermally oxidised samples with

particular focus on automated porosity classification[124]. More recent

work has used synchrotron tomography with phase contrast

reconstruction algorithms to investigate crystal strain due to thermal

expansion[125], phase contrast techniques are suited to graphite

experiments due to the low absorption co-efficient of graphite[126].

3.1.2.2. Digital Image Correlation

DIC is an image analysis technique used to track movement of features

between images. The technique was developed during the 1980’s when

image correlation found use for engineering applications in particular

Page 79: Using Intercalation to Simulate Irradiation Damage of

79

stress analysis in solid mechanics and particle image velocimetry in fluid

mechanics.

By measuring the movement of pixel patterns between two images

displacement vectors can be calculated and from this microstructural

dimensional change can be calculated [127]. In the most basic sense

this is achieved by sectioning an image and tracking the movement of

the pixel patterns and calculating the associated displacement vector.

For a good quality analysis there must be sufficient texture within the

reference window to track motion. For unmodified materials this

means increasing the size of the window. However if the window is too

large motion cannot be tracked within that window. Errors can be

further reduced by ensuring the movement between each image is as

small as possible[128].

Two Dimensional DIC was first applied to fracture mechanics problems

and there have been a number of such studies devoted to

understanding crack propagation in nuclear graphite[129-131]. More

recently three dimensional DIC has been carried out looking at strains

induced by thermal expansion of nuclear graphite. This work focused

on understanding which regions within the heterogeneous

microstructure contributed to expansion. Algorithms were written to

extract local strains from the displacement vectors measured by DIC.

The code removes rigid body motion, rotational vectors which are

induced by small rotational displacements between different

tomography scans and finally radial expansion due to the bulk

expansion. By removing the bulk displacements the remaining

displacement vectors describe the local strains and as such allow an

Page 80: Using Intercalation to Simulate Irradiation Damage of

80

understanding of how microstructural features affect microstructural

strains. The work proved very effective in characterising local

displacements[125].

3.1.3. Laser Ultrasonics

Ultrasound refers to pressure waves with a frequency in the range of

20kHz to 2GHz and are useful for non destructive engineering

applications such as crack detection and property measurements.

Ultrasonics are commonly used for the measurement of Young’s

modulus in nuclear graphite. This section will introduce the standard

test method to perform Dynamic Young’s Modulus (DYM)

measurements and go on to describe the modifications employed to

apply the technique to the measurement of brominated graphite.

3.1.3.1. Ultrasonics

Equations defining DYM measurements differ to classical modulus

definitions because the strains induced are small. Therefore the

definition is derived from equations of motion, a full derivation can be

found in Timoshenko’s Theory of Elasticity[132] and the result is given

in Equation 3.2.

)1(

)21)(1(2

vE Equation 3.2

where E is the DYM, ρ is the sample density, η is Poisson’s ratio and ν is

the ultrasonic velocity. Though ultrasonic waves propagate through

elastic media in three modes, shear waves, longitudinal waves and

Page 81: Using Intercalation to Simulate Irradiation Damage of

81

surface waves, for DYM measurements only the longitudinal velocity is

required[133].

3.1.3.2. ASTM standard C 769-09

The experimental method to measure the DYM of nuclear graphite is

defined by ASTM standard C 769-09. Using piezo transducers the

experimental setup is outlined in Figure 3.4. The experimental

apparatus measures the time an ultrasonic pulse takes to travel

through a sample. This standard is reported to give a result to within

10% of the result provided by other methods such as measuring

fundamental frequencies[134].

Figure 3.4 Schematic of ASTM C769 experimental setup[134]

The driving circuit should be capable of generating frequencies from 0.5

to 2.6MHz. Samples must be straight with a uniform cross section. The

end faces must be perpendicular to the cylindrical surface within to

0.125mm and the sample must be long in comparison to the probing

wavelength. The weight and dimensions of the sample are to be

weighed to 0.5% accuracy.

Page 82: Using Intercalation to Simulate Irradiation Damage of

82

3.1.3.3. Laser Induced Ultrasound

As the principle focus of this research is aimed at understanding

property changes of brominated samples a technique was developed to

operate in harsh corrosive environments. To this end the use of laser

ultrasonics was utilised as a non contact method. Laser ultrasound has

a number of advantages over differing contact methods as the

measured sample is not held in a pre-stressed state as well as problems

associated with transducer contact couplants such as difficulties in

ensuring optimum contact conditions. Most importantly though for this

research the method can be applied for inaccessible samples in difficult

environments. Previous work has used lasers to induce ultrasound in

graphite[135].

It sounds counter intuitive that light can induce stress waves but this

can be achieved using coherent light from lasers. The non reflected

portion of light energy has a number of effects on the sample

depending on the wavelength and energy. At low power heating and

thermal and elastic waves are produced while at higher powers there

may be melting, plastic deformation and crack formation. For Non

Destructive Testing (NDT) low powered lasers are therefore required.

An elastic wave is generated by the sample absorbing electromagnetic

radiation which heats up a small region of the sample. This in turn

causes rapid thermal expansion of a small region which causes an

elastic wave to propagate through the solid as shown in Figure 3.5.

Page 83: Using Intercalation to Simulate Irradiation Damage of

83

Figure 3.5 Schematic of ultrasonic wave generation by laser [136]

The ideal laser for generation of elastic waves is pulsed as this produces

rapid temperature changes in the sample for short periods which

generates thermoelastic stresses rather than heating of the sample. The

pulse duration should be of a magnitude such that the induced stress

waves have a suitable wavelength. A standard 20ns pulse induces a

20MHz ultrasonic wave in steel [136].

3.1.3.4. Laser Interferometry

Lasers can also be used to measure surface displacements with high

accuracy. This can be achieved by two different methods, interference

between an incident and reference beam providing instantaneous

displacement measurements or a second arrangement measuring the

change in frequency of the reflected incident beam gives good velocity

measurement. As non contact methods lasers have the advantage over

standard piezo measurements as the laser does not interact with the

ultrasonic field.

Page 84: Using Intercalation to Simulate Irradiation Damage of

84

The basic interferometer setup used is shown in Figure 3.6. A laser

beam is fired at a beam splitter. Part of the beam is reflected off a fixed

mirror and the remaining beam is reflected off the sample surface. The

two beams are recombined and then directed to the detector surface.

As the sample surface moves the path length of the measuring wave

changes. This causes a change in the interference of the two light

beams which is measured by the detector [136].

To reduce errors associated with a laser interferometery measurements

there are a number of important factors to consider with the

experimental setup. Firstly ensure external vibrations are eliminated,

this is most easily achieved using an air table.

If the sample is measured through glass it is important to have the

window at an angle off parallel to the sample surface to reflect this

portion of the signal away from the vibrometer. However the sample

surface should be as close to perpendicular as possible as this

minimises speckle interference [137]. Speckle interference arrises from

a surface roughness on the scale of the impinging wave causing the

reflected waves to interfere. The interference pattern changes as the

surface vibrates causing the signal to fluctuate, this problem can be

reduced by slight defocusing of the beam[138].

Page 85: Using Intercalation to Simulate Irradiation Damage of

85

Figure 3.6 Michellson interferometer

3.1.4. X-ray Diffraction

XRD is a technique used to determine the crystal structure of materials

and is used in this study to characterise different graphitic

microstructures. This study uses two x-ray techniques for

characterisation, powder diffraction and textural analysis.

Powder diffraction is a very powerful technique and can be used to

establish many aspects of crystal structure including lattice parameters,

crystal size, strains and dislocation densities. During this research

various graphite grades have been measured using XRD in order to

calculate the lattice parameters and crystal size

Textural analysis is a technique used to determine the preferred

orientation within a sample. This is important to this study to gain an

insight into how the initial arrangement of crystals affected subsequent

dimensional changes under intercalation conditions[139].

Laser

Detector

Fixed Mirror

Moving

Mirror

Page 86: Using Intercalation to Simulate Irradiation Damage of

86

All XRD measurements conducted for this research were carried out

using the Philips X’Pert modular diffractometer shown in Figure 3.7.

This is a very versatile piece of equipment and can be setup for powder

diffraction measurements and texture measurements. The x-rays are

produced from a cobalt anode with a wavelength of 1.7902Å.

Figure 3.7 Philips X’pert modular diffractometer

3.1.4.1. Lattice Parameters

X-rays have wavelengths of a similar magnitude to interatomic spacing

and are therefore diffracted. The interaction of x-rays and atoms is

complex, however the problem has been simplified by William Bragg

[140] who proposed the schematic shown Figure 3.8 which forms the

basis of all XRD.

Figure 3.8 Schematic representation of diffraction by a crystal[39, 141]

Incident

rays

Reflected

rays

Page 87: Using Intercalation to Simulate Irradiation Damage of

87

The schematic shown in Figure 3.8 describes the Bragg equation where

the lattice spacing of hkl planes can be calculated by measuring the

angle of peak intensity known as the Bragg angle, Equation 3.3.

Equation 3.3

where dhkl is the lattice spacing for the hkl lattice plane, lambda is the

wavelength of incident x-ray and sinθ is the angle of the reflected x-ray

[141]. Figure 3.9 details some of the important hkl planes in graphite.

Figure 3.9 Schematic of hkl planes[20]

3.1.4.2. Crystallite Size

The size of crystals can be determined to a first approximation using the

Scherrer equation [142]. The Scherrer equation given in Equation 3.4

gives the lower bound for crystal size. The crystal thickness (t) is given

by relating the incident wavelength to the broadening of a peak in the

xrd spectra. The peak broadening is quantified by measuring the peak

full width at half maximum (FWHM).

Page 88: Using Intercalation to Simulate Irradiation Damage of

88

Peak broadening comes about from x-rays reflecting at slightly differing

Bragg angles which changes the destructive interference pattern. Peak

broadening appears due to a number of reasons; instrumental

broadening, strain broadening as well as size broadening. Instrumental

broadening can be removed by appropriate use of a standard, however

size and strain broadening are more difficult to separate. Size and strain

broadening are closely linked and therefore the Scherrer equation

should be used as a lower bound[141] for crystal size. This work utilised

a silicon standard in order to determine the instrumental broadening.

BBt

cos

9.0

Equation 3.4

3.1.4.3. Texture Analysis

The preferred crystal orientation of crystals is different for different

graphite grades and is strongly dependant on the manufacturing

process and raw materials used. The orientation can be measured by

XRD textural analysis, where the sample is rotated with respect to the

incident beam. The XRD is focused on a particular peak and the

variation in peak intensity is measured. A high peak intensity means

that a large number of crystal planes are orientated perpendicular to

this direction where as a low intensity means there are few planes in

the corresponding orientation[143]. Bacon uses this method to derive

an anisotropy factor for nuclear graphite, Equation 3.5. By focussing on

the variation in intensity of the 002 plane[144] he predicts the equation

could be used to predict the performance of graphite in irradiation

conditions.

Page 89: Using Intercalation to Simulate Irradiation Damage of

89

2

0

2

2

0

3

..

..

.sin.cos).(

.sin).(

2

1

..

..

dI

dI

CD

CD

GW

GA Equation 3.5

Here Φ is the azimuth angle and I(Φ) is the intensity at that angle and

D.C.A.G./ D.C.W.G. is the predicted Bacon Anisotropy Factor (BAF).

3.1.4.4. Pycnometry

Pycnometry was carried out using a Micrometrics AccuPyc 1340

analysis system shown in Figure 3.10. Pycnometry is a technique which

measures the volume of a sample. This is achieved by placing a sample

in a chamber at ambient temperature and pressure, and, upon closing

the chamber, the pressure is increased. Adjacent to the sample cell is

another cell of a calibrated volume, a valve is released allowing the gas

to escape in to the second chamber. By measuring the subsequent

pressure the volume of the sample and the closed porosity can be

calculated with Equation 3.6[145].

12

1

g

g

EXPCELLSAMP

P

P

VVV

Equation 3.6

where VSAMP, VCELL and VEXP are the volume of the sample, sample cell

and expansion cell respectively. P1g is the chamber pressure at the start

Page 90: Using Intercalation to Simulate Irradiation Damage of

90

of the experiment i.e. ambient and P2g is the pressure once the

expansion chamber has been opened.

Figure 3.10 Micrometrics pycnometer

3.2. Experimental Rigs

3.2.1. Design of experimental rig to measure dimensional change of

bulk material

Experiments to measure the dimensional change induced by

intercalation of bromine were carried out using an experimental rig

based on work by Brocklehurst[63]. The vast majority of the design

work was carried out by Miss Perrin as part of her MSc thesis[146].

Figure 3.11 shows the resulting experimental setup.

The main design aspect to this rig is the design of the spring. Previous

work of this type have used quartz springs[63, 118, 147] as quartz will

remain chemically and structurally uncompromised in the strong

oxidizing atmosphere. Unfortunately it was not possible to obtain a

suitable quartz spring, therefore it was decided that tantalum, another

material unaffected by bromine, would be used. The tantalum springs

Page 91: Using Intercalation to Simulate Irradiation Damage of

91

were made by hand and calculations showed that a spring radius of

25mm with a wire diameter of 0.6mm and 21 turns would provide the

necessary deflection across the anticipated mass change[146].

For initial experimentation with the experimental set up, a flask heater

was used to control the temperature of the bromine. This was found to

be of little advantage experimentally whilst increasing the risks

associated with the experiment and was therefore disregarded for data

collection runs.

Figure 3.11 Dimensional Change Rig

3.2.2. Design of experimental rig to measure microstructural

dimensional change

An experimental rig was built to enable tomography scans to be made

of brominated graphite samples in order to gain an understanding of

dimensional change within the heterogeneous microstructure of

nuclear graphite. The first point to consider was the selection of a

suitable material for the sample container. For safety the rig must be

Page 92: Using Intercalation to Simulate Irradiation Damage of

92

stable in a heavily oxidising environment; however for the best results

the rig must absorb as few x-ray photons as possible. A rig which

absorbs a high number of photons would result in noisy data at best

and no data at worst.

With this in mind there were two candidate materials for the rig, glass

and PTFE. The energy range of photons of the tomography beam line at

the Swiss Light Source (SLS) are in the 10keV range[148]. At this energy

the absorption coefficients of glass and PTFE are 1.71 cm2/g and 0.68

cm2/g respectively, Figure 3.12. The attenuation coefficient is defined in

Equation 3.7.

)/ln(/ 0

1 IIx Equation 3.7

where x = ρt is the density times the sample thickness. The thinnest

glass available for the rig was 1mm thick. Although thin layers of PTFE

have poor structural integrity it is possible to lend support by applying a

thin coating of PTFE to an aluminium support. The absorption

coefficient of aluminium at 10KeV is relatively high, 2.62 cm2/g. It was

possible to obtain aluminium tubes of 0.5mm thickness and apply a

PTFE coating of 25μm.

Page 93: Using Intercalation to Simulate Irradiation Damage of

93

X-ray attenuation coefficients of potential rig materials

1.00E-02

1.00E-01

1.00E+00

1.00E+01

1.00E+02

1.00E+03

1.00E+04

0.00 0.01 0.10 1.00 10.00 100.00

Photon Energy (MeV)

u/p

(c

m^

2/g

)

Aluminium

PTFE

Glass

Figure 3.12 X-ray attenuation coefficients of candidate rig materials[149]

Table 3.1 shows the theoretical loss of beam intensity after travelling

though the potential candidate materials. The logarithmic relationship

between material thickness and attenuation means that it is more

important to use the thinnest materials possible rather than those with

the lowest attenuation coefficient and therefore aluminium with a PTFE

coating was used.

Table 3.1 Theoretical loss of x-ray intensity due to rig

3.

Attenuation coefficient

(cm2/g)

Material thickness

(mm)

Density

(g/cm2) I0/I

PTFE 0.68 0.5 2.2 2.11

Aluminium 2.62 1 2.7 1181

Glass 1.71 3 2.23 92958

The second aspect to consider was the sample dimensions. It was

considered important to obtain the highest resolution possible in order

Page 94: Using Intercalation to Simulate Irradiation Damage of

94

that very small strains may be measured. The available options are

given in Table 3.2. The best option would be 0.37µm, unfortunately, it is

not possible to machine samples to such high specs. Therefore, a

compromise was reached with the area of interest of the sample being

1mmØ. A base was added to the sample to provide stability as shown in

Figure 3.13.

Table 3.2 Microscope objectives available on TOMCAT beamline[148]

Magnification Field of View

(mm2) Pixel Size

(µm2)

1.25 12.1 x 12.1 5.92 x 5.92

2 7.5 x 7.5 3.7 x 3.7

4 3.7 x 3.7 1.85 x 1.85

10 1.5 x 1.5 0.74 x 0.74

20 0.75 x 0.75 0.37 x 0.37

Figure 3.13 Tomography sample

To maximize the use of the allocated beamtime it was decided to stack

samples in the rig. The sample stage at TOMCAT has 25mm vertical

motion; therefore all samples sat within this range. Below the samples

a bromine containment vessel was located. The bromine vessel allowed

Page 95: Using Intercalation to Simulate Irradiation Damage of

95

the bromine to be released at a time of the experimentalists’ choosing,

allowing tomographic scans of unbrominated specimens to be taken.

To allow the release of bromine to be controlled by the experimentalist

a glass vial with a stopcock was specifically manufactured for the task.

The vessel height had a strict dimensional specification of 55mm as this

was the height required for the sample when the sample stage was at

its lowest. The final design is shown in Figure 3.14. To ensure the design

did not leak the PTFE based grease Lox-8 was used to seal all joints.

Figure 3.14 Experimental rig for tomographic scans of bromine intercalated graphite

3.2.3. Design of experimental rig to measure Young’s modulus change

of brominated graphite

Page 96: Using Intercalation to Simulate Irradiation Damage of

96

To measure the change in Young’s modulus of brominated graphite

samples it was necessary to construct a rig which would remain stable

in a highly oxidising environment whilst providing good access to the

lasers used to induce ultrasonic waves and measure their arrival on the

opposing sample surfaces. From the experience gained in previous

experiments it was decided that the bulk of the rig would be made from

PTFE and that Lox-8 grease would be used to seal the rig.

The first factor to consider was the type of laser used to induce

ultrasound. The shorter the wavelength of the impact laser the higher

the proportion of energy that is absorbed by the sample surface and

therefore the more efficient the generation of ultrasound. This is

because the reflectivity of a surface is related to the wave length of the

incoming photons [136]. Care must be taken however as shorter

wavelengths will induce ablation at lower fluence [150]. The laser

eventually chosen to induce the ultrasound was a Nd:YAG laser which

produces photons with a wavelength of 1064nm.

A coating was applied to the Nd:YAG access window to reduce the

reflection and ensure the maximum possible amount of energy reached

the sample surface. Argon coated windows for the rig were purchased

from Edmund Optics. Figure 3.15 shows that the reflectivity of Argon

coated glass for photons with a wave length of 1064nm is close to zero.

The arrival of the sonic pulse on the opposing sample surface to that of

the impact laser was measured using a HeNe laser. HeNe lasers are

often used in laser vibrometry as they have good coherence properties

and well defined wavelengths[136]. The wave length of a HeNe laser is

633 nm. It was not possible to obtain a coated window exactly to suit

Page 97: Using Intercalation to Simulate Irradiation Damage of

97

this wavelength, though Edmund Optics VIS-NIR coated glass is

reasonably good with around 1.4% reflectivity, Figure 3.16.

Figure 3.15 Reflectivity of argon coated glass[151]

Figure 3.16 Reflectivity of VIS-NIR coated glass[151]

To reduce the number of bespoke components required for the rig, the

bromine containment vessel previously used for the tomographic rig

was utilised. Two access ports were added complete with quartz

windows to make it easy to view the sample whilst aligning the lasers.

An attempt was made to collect images for digital image correlation

through these windows. However as the chamber filled with bromine it

Page 98: Using Intercalation to Simulate Irradiation Damage of

98

became impossible to obtain images of a suitable quality. Ports are

sealed with Lox-8 grease and PTFE seals. Figure 3.17 shows the

exploded diagram of the bromination chamber.

Figure 3.17 Exploded diagram of cell to measure Young's modulus of brominated graphite

The strain of the sample was measured using a laser displacement

detector measuring the displacement of one end of the sample surface.

This requires one end of the sample to be fixed. The sample holder for

the polycrystalline samples was a PTFE base with a tantalum wire

triangle protruding out. This allowed easy access for the impact laser

and ensured that the sample expanded in the opposing direction,

Figure 3.18. The sample was allowed to slightly overhang the sample

edge so that it rested with one end close to the optical window.

Page 99: Using Intercalation to Simulate Irradiation Damage of

99

Figure 3.18 Sample holder for polycrystalline graphite

3.3. Conclusions

This chapter details the experimental techniques and rigs that were

used throughout this study.

Experimental techniques covered are the use of sorption balances for

bulk dimensional change experiments. Tomography and DIC for

microstructural dimensional change experiments. Dimensional change

is covered in Chapter 5. The standards and theory behind DYM

measurements are discussed which are used in Chapter 6 to investigate

changes in Young’s modulus. XRD and pycnometry are covered which

have been used in Chapter 4 for microstructural characterisation of the

graphites used in this study.

A rig has been developed based upon Brocklehurst’s rig used in for his

studies into brominated graphite [28]. The rig required some

modifications due to the current availability of components. A second

rig was designed to investigate microstructural changes due to

bromination using tomography. A third rig was designed to measure

the changes in Young’s modulus that are induced by bromination.

Page 100: Using Intercalation to Simulate Irradiation Damage of

100

4. Characterisation of nuclear graphite

Over the course of this study a number of different grades of graphite

have been examined, all of which are currently or have been used in

the past in nuclear reactors. This chapter details a number of

experiments carried out to characterise the microstructure of these

graphites. The chapter concludes with background detail on the

graphites and a summary of the main characteristic features of the

graphites.

4.1. Characterisation Results

4.1.1. Polarised optical microscopy

Figure 4.1 presents optical micrographs of the three nuclear graphites

used in this study, PGA, Pechinay and Gilsocarbon. As discussed in

Chapter 1 there are common microstructural features seen in nuclear

graphite and these are highlighted in the images.

PGA and Pechiney graphites both have pointed ‘needle like’ filler

particles whereas Gilsocarbon has large spherical filler particles. All

three graphites exhibit the different types of porosity visible at this

level of magnification, gas evolution pores and calcination cracks. The

polarised light highlights the disorder of crystallites in the binder matrix

and the order within the filler particles.

Page 101: Using Intercalation to Simulate Irradiation Damage of

101

a) Pile Grade A

b) Pechiney

Binder Matrix

Filler

Particle

Calcination crack

Gas evolution pore

Filler

Particle Binder

Matrix Calcination crack

Gas evolution pore

Page 102: Using Intercalation to Simulate Irradiation Damage of

102

c) Gilsocarbon

Figure 4.1 Optical mircographs of nuclear grade graphite

4.1.2. Pycnometry

The pycnometry samples used were 10Ø x 15mm. They were

manufactured by lathe, cleaned in an ultrasonic bath and then left to

dry for 2 weeks. The volume of the sample was then calculated using

digital callipers correct to two decimal places. The mass of the sample

was measured on a 2 decimal place electronic balance.

The open porosity is found as the difference in bulk volume measured

by the pycnometer and the bulk volume measured using callipers. The

closed volume is found by taking the density of graphite as

2.25g/cm3[60], the density of a graphite crystal, and comparing it to the

measured mass of the sample using Equation 4. 1[152].

Calcination crack

Gas evolution pore

Binder Matrix Filler Particle

Page 103: Using Intercalation to Simulate Irradiation Damage of

103

1

1

100(%) C

OpenPBulk

sample

BClosedPVV

mV

Equation 4. 1

where VClosedP, VBulk and VOpenP are the volume of closed porosity, bulk

volume and open porosity as measured by the pycnometer. ρB is the

bulk density and ρc in the crystal density.

The results shown in Table 4.1 are comparable with Standring’s results

for PGA[152]. Standring found an open pore volume of 20.1% and a

closed pore volume of 5.5%. The results show that PGA has the largest

pore volume but the lowest closed pore volume. Gilsocarbon and

Pechiney which have undergone reimpregnation during manufacture

have low overall porosity but a higher volume of closed porosity

Table 4.1 Pycnometry data for selected nuclear graphites

4.1.3. Powder Diffraction

The XRD was carried out on samples dimensioned 20 x 20 x 5mm

samples. These are relatively large sample dimensions and ensured high

Page 104: Using Intercalation to Simulate Irradiation Damage of

104

quality data collection by limiting unwanted reflections. The data was

collected at the highest resolution time allowed which in this case was

0.025 degrees with a time step of 10 seconds. This gave a minimum of

1000 counts for the smallest peak of interest, the {110}. The other main

peak of interest is the {002} has a significantly stronger reflection.

The XRD powder diffraction data, shown in Figure 4.2, Figure 4.3, and

Figure 4.4, was used to gain an understanding of the crystal structure.

First the d spacings were calculated using Braggs equation. The results

shown in Table 4.2 show that PGA has the best layer plane stacking

indicated by the low Lc value 3.91 Å, Gilsocarbon has the widest Lc

spacing of 3.94 Å. All the graphites have very similar values for a

spacing of 1.43Å.

Page 105: Using Intercalation to Simulate Irradiation Damage of

105

0 20 40 60 80 100 120

Bragg angle (2θ)

Inte

nsit

y (a

rbit

rary

uni

ts)

{002}

{100}

{101}

{004} {110} {112} {114}{006}

Figure 4.2 XRD powder diffraction spectra of PGA

0 20 40 60 80 100 120

Bragg angle (2θ)

Inte

nsit

y (a

rbit

rary

uni

ts)

{002}

{100}

{101}

{004} {110} {112} {114}{006}

Figure 4.3 XRD powder diffraction spectra of Pechinay Graphite

Page 106: Using Intercalation to Simulate Irradiation Damage of

106

0 20 40 60 80 100 120

Bragg angle (2θ)

Inte

nsit

y (a

rbit

rary

uni

ts)

{002}

{100}

{101}

{004} {110} {112} {114}{006}

Figure 4.4 XRD powder diffraction spectra of Gilsocarbon

Table 4.2 Powder diffraction data

To find crystal size is slightly more involved. The size contribution to the

full width at half maximum (FWHM), B struct, is found by subtracting

the standard FWHM from the sample FWHM. The FWHM describes the

width of a peak at half of its maximum height as shown in Figure 4.5. To

calculate the crystal sizes, a silicon standard was used to remove the

instrumental broadening from the measurement.

Page 107: Using Intercalation to Simulate Irradiation Damage of

107

Figure 4.5 Image detailing the Full Width at Half Maximum[153]

The results given in Table 4.2 show that PGA has significantly larger

crystals than the other graphites (425 x 369 Å). This figure is almost

twice that of Pechinay and Gilsocarbon with dimensions of 293 x 283 Å

and 235 x 241 Å respectively.

4.1.4. Textural Analysis

The XRD was carried out on samples dimensioned 20 x 20 x 5mm. The

samples were positioned such that the Against Grain (AG) direction ran

parallel to the sample surface and pointed to an azimuth angle of 0

degrees. The data was collected for the {002} and {110} peaks. The

{002} gives the orientation of the layer planes, the {110} gives the

orientation of the a-planes.

The measurements were carried out with a generator voltage of 40kV

and a tube current of 40mA. Figure 4.6 to Figure 4.8 show the raw pole

figures. Before any meaningful analysis could be achieved, correction

Page 108: Using Intercalation to Simulate Irradiation Damage of

108

factors had to be applied. First a background correction must be

applied. The background error can come from a number of sources but

is particularly sensitive to changes in the tilt angle. By measuring a

nearby section of the XRD spectra which is known to contain no peaks

and is far enough away from the measured peak to allow for peak

broadening at high tilt angles, it is possible to adjust for changes to the

background[143].

The next correction to apply is the defocusing factor. This arises from

geometric considerations of the experimental setup. It is therefore

possible to correct for this by using a geometric correction function.

However, it is often better to use empirical data to obtain the

correction factor. This is achieved using a totally random sample as a

base line. As it was not possible to obtain a standard for this

experiment, the Gilsocarbon pole figure shown in Figure 4.8 was used.

The Gilsocarbon pole figure shows very random orientation. It is

sometimes necessary to also apply an absorption correction factor to

the data. The samples used for this experiment were thick enough, and

therefore this was not necessary [143].

The PGA pole figure shown in Figure 4.6, displays a strong orientation in

the {002} plane along the central axis of the pole figure. The {110}

Page 109: Using Intercalation to Simulate Irradiation Damage of

109

shows strong orientation perpendicular to the axis where strong

orientation is seen in the {002} pole. The Gilsocarbon pole figure shown

in Figure 4.8, demonstrates almost no orientation and the intensity of

reflection shows almost no variation with azimuth angle. This is

matched by very little intensity variation in the {110} pole figure. This

can be confirmed quantitatively as the Bacon anisotropy factors given

in Table 4.3. Further analysis of texture data is carried out in the

following chapter.

Table 4.3 Bacon Anisotropy Factors

PGA Pechiney Gilsocarbon

Bacon Anisotropy Factor 1.4432 1.1908 1.0216

Page 110: Using Intercalation to Simulate Irradiation Damage of

110

Figure 4.6 PGA pole figure

Page 111: Using Intercalation to Simulate Irradiation Damage of

111

Figure 4.7 Pechiney pole figure

Page 112: Using Intercalation to Simulate Irradiation Damage of

112

Figure 4.8 Gilsocarbon pole figure

Page 113: Using Intercalation to Simulate Irradiation Damage of

113

4.2. Graphite Grades

4.2.1. Pile Grade A graphite

Pile Grade A (PGA) graphite was used in the British Magnox program.

This graphite was manufactured to a specification dictated by reactor

physicists in order to obtain good moderation properties (very low

absorption cross section ≤ 4.1millibarns and density ≥1.7g/cm3), which

required a low degree of impurity concentrations (ppm / bpm). The

result was very pure graphite with a large grain size and high degree of

crystallinity. The base materials are needle coke filler particles derived

from petroleum coke and a “low ash” coal tar binder pitch. This

graphite was formed by extrusion and this in combination with the

large grain size produced a highly anisotropic material.

Pile Grade A is characterised by long needle shaped filler coke particles

Figure 4.1(a). The filler particles contain calcination cracks and are

believed to form during the calcining process. In addition there are a

whole range of small cracks, nanometres in width by micrometres in

length. These are attributed to the large difference in the crystal

coefficient of thermal expansion perpendicular and parallel to the basal

planes. This difference leads to micro-cracking on cooling from the

graphitisation temperature [19]. These nanocracks tend to run parallel

to the crystal basal layer planes and are generally closed porosity (i.e.

not accessible to surrounding gas or air). As PGA is formed by an

extrusion process the needle type filler particles tend to align

themselves with the direction of extrusion, this direction is known as

With Grain (WG). The filler particles are surrounded by a binder matrix

Page 114: Using Intercalation to Simulate Irradiation Damage of

114

which contains “flour”, that is small ground petroleum coke particles

randomly orientated mixed with coal tar pitch. The matrix is permeated

by gas evolution pores which form during the baking process[154, 155]

which tend to form open porosity. PGA has the largest open pore

volume of all the graphite grades studied in this work, measured by the

author as 23% , this is just above the quoted literature values around

19.8%[35]. The graphite also has the largest crystals size measured by

XRD as 42.5 x 36.9nm and the highest Bacon anisotropy factor 1.44.

4.2.2. Pechiney Graphite

Pechinay graphite was used in the French Magnox reactors[156]and is

comparable to PGA. It is a very pure extruded graphite with needle like

filler particles. The measured open pore volume is around 15%, the

crystallites are also smaller than PGA around 29.3 x 28.3nm. The

graphite is also less anisotropic than PGA with an anisotropy ratio of

1.19.

4.2.3. Gilsocarbon

Gilsocarbon was a grade manufactured by two graphite companies

Anglo Great Lakes (AGL) and British Acheson Electrodes Ltd (BAEL). It

was used in the British AGR reactors, Germanys Thorium High

Temperature Reactor (THTR) and as fuel supports in some of the French

Magnox reactors. The grade was developed for the more challenging

operating conditions of the AGR which were more highly rated than the

Magnox reactors. In particular a graphite was required that was

isotropic, more resistant to oxidation and with increased dimensional

stability under irradiation.

Page 115: Using Intercalation to Simulate Irradiation Damage of

115

The graphite is based on the coke obtained on refining a natural asphalt

“Gilsonite” mined in Utah[157, 158]. The coke particles are spherical

with a lamina layered structure often referred to as ‘onion like’ in which

the crystallites tend to be orientated perpendicular to the sphere

radius. This gives a bulk material with near isotropic properties, with a

Bacon anisotropy factor of 1.02. Gilsocarbon has been measured by the

author to have an open porosity of 11% which agrees well with values

in the literature[35] and small crystallites compared to the anisotropic

graphite grades.

4.3. Conclusions

This chapter details microstructural characterisation experiments

carried out on the grades of nuclear graphite used for investigations

into the effect of bromination on changes to material properties. Three

graphites have been used PGA, Pechiney, and Gilsocarbon. PGA and

Pechiney are both highly anisotropic graphites as highlighted by their

high values for the BAF calculated from pole figure measurements. Both

PGA and Pechiney have needle like filler particles. Gilsocarbon is a near

isotropic graphite with spherical filler particles. PGA has the largest

volume of porosity within its structure followed by Pechiney and

Gilsocarbon. PGA is made of the largest crystallites with the smallest d

spacing whilst Gilsocarbon has the smallest crystallites and the largest d

spacing’s.

Page 116: Using Intercalation to Simulate Irradiation Damage of

116

5. Dimensional Change

Graphite is subjected to extreme environments in nuclear

applications, with variables such as neutron flux and temperature

inducing anisotropic dimensional change in the crystals as has been

described in Chapter 1. Problems arising from dimensional change

effects are exacerbated by variations in the flux intensity around a

reactor core. The flux profile seen by a component is not uniform

across a reactor, or even a component and this causes variations in

the local dimensional change rates[34]. The different dimensional

change rates can lead to brick deformations such as bowing and

barrelling examples of which are as shown in Figure 5.1.

a)

b)

c)

Figure 5.1 Brick deformations caused by non uniform flux profiles a) bowing b) barrelling c) wheat sheafing

Bowing is caused by a flux profile across the component causing a

difference in expansion across the brick. The flux profile can be

present because of a number of factors including proximity to

absorbers, empty fuel channels or side reflectors. Bowing can cause

Page 117: Using Intercalation to Simulate Irradiation Damage of

117

instability in the reactor core as wedge shaped gaps open between

bricks in a column.

Barrelling and wheat sheafing are caused by the radial flux profile of

a brick. The flux is higher towards the centre of a brick where it is

closer to the fuel. Initially there are tensile stresses around the bore

of the brick and compressive stresses towards the outer edges. These

stresses combined with the lack of axial restraint at the bricks ends

cause the brick to barrel. Wheat sheafing occurs later in the reactor

lifecycle as turnaround occurs earlier in crystals around the bore of

the brick creating a reversal in the radial stress profile in the brick.

Structural features such as keyway slots introduce sharp radii into

graphite components. The effect of irradiation induced strains at the

resulting stress concentration points can cause components to

fracture, gradually reducing the structural integrity of the core.

Dimensional change is therefore a key property in determining the

operational lifetime of reactors, should the effect become too

considerable the performance, structural integrity and ultimately the

safe operation of the reactor are compromised.

This chapter aims to gain an understanding into how changes in

crystal volume directly affect the graphite microstructure by using

bromine intercalation to simulate irradiation damage. The effects of

bromine intercalation on HOPG and polycrystalline graphites are

measured using Brocklehurst’s methods[63]. Here the work is

Page 118: Using Intercalation to Simulate Irradiation Damage of

118

furthered by using x-ray tomography and modelling to gain an

understanding of microstructural changes in damaged graphite. High

resolution tomography is a powerful technique which is used to gain

an understanding of how microstructural features affect the

intercalation of bromine and how the development of the

intercalation compound drives dimensional change of the bulk

structure. The modelling based on the techniques of Sutton and

Howard[46] provides an insight into processes which are occurring

below the resolution of tomography.

5.1. Dimensional Change of Nuclear Grade Graphite by Bromine

Intercalation

Using the experimental rig shown in Figure 5.2, it was possible to

understand bromination on the bulk structure of graphite. The

bromine used was analytical reagent grade, a grade that is

sufficiently pure for chemical analysis but is by no means the purest

bromine available. The bromine is allowed to evaporate at room

temperature and pressure in a sealed glass column which will

equilibrate with a vapour pressure of 223 hPa[159].

Page 119: Using Intercalation to Simulate Irradiation Damage of

119

Figure 5.2 McBain Sorption Balance[63]

5.1.1. Calibration of the spring for sorption balance

As described in Section 3.2.1 the experiment uses a McBain Spring

Balance[118] to measure the sorption of the sample. This is achieved

by hanging a graphite sample from a spring. The springs were

designed such that the anticipated change in mass would allow the

fullest range of the travelling micrometer to be used[146].

The springs were calibrated hanging weights of known mass off the

end of the spring. The weights were measured on a digital balance

accurate to milligrams. The spring was hung from a clamp stand and

a small piece of masking tape was stuck to the bottom of the spring

as a marker. The height of the marker was measured with a travelling

micrometer. The resulting curves given in Figure 5.3 and Figure 5.4

Travelling Microscope

Graphite

Sample

Bromine

Spring

Page 120: Using Intercalation to Simulate Irradiation Damage of

120

show the springs to have a very linear response over the mass range

of interest with R2 values of 0.9976 and 0.9949. The curve for the

polycrystalline experiment does not lie completely within the

expected error bounds. Therefore, there has been an

underestimation of some associated errors namely the linearity of

the spring. This is most likely due to the fact that the spring was

handmade and so the coils won’t all be parallel.

Figure 5.3 Calibration of spring for weighing polycrystalline samples

Figure 5.4 Calibration of spring for weighing HOPG samples

Page 121: Using Intercalation to Simulate Irradiation Damage of

121

5.1.2. Dimensional change experiments

The dimensional change experiments were carried out using HOPG

and polycrystalline graphites. HOPG was used as an approximation to

single crystal graphite. The polycrystalline graphites used were PGA,

Pechiney and Gilsocarbon chosen for differences in their respective

microstructures.

Measurements are carried out on the samples prior to starting the

experiment. The mass of the sample is measured on a digital balance

accurate to milligrams and the dimensions are measured with

callipers. The graphite sample is then attached to the spring and

carefully placed in the glass column. Care is taken to ensure that this

is done slowly to prevent the sample oscillating too much. As soon as

any oscillations have stopped two height measurements, highlighted

in Figure 5.5 , are taken with a travelling microscope. These

measurements are repeated at time intervals until the end of the

experiment.

Figure 5.5 Schematic of McBain Spring Balance

h1

h2

1500mm

0mm

Orientation of

travelling

micrometer scale

Page 122: Using Intercalation to Simulate Irradiation Damage of

122

The dimensional change of the sample is calculated by comparing the

dimensions of the original sample to those of the sample at each

time increment using Equation 5.1, where h1 and h2 are the

measurements taken at each time interval and h1t0 and h2t0 are the

initial height measurements taken. For the first measurement, the

height of the sample is compared to the length measurement carried

out with callipers. This showed there was no measurable dimensional

change occurring between the sample entering the bromine chamber

and the spring stopping oscillating, Therefore h1t0 and h2t0 are taken

as the length measurements at t0.

0

0

21

2121..

tto

tto

hh

hhhhCD

Equation 5.1

The amount of bromine absorption that has occurred is calculated as

the molar ratio of bromine and carbon present in the sample.

Equation 5.2 gives the equation used to make the calculation. Here k

is the spring constant measured in Figure 5.2 or Figure 5.3, M(C) and

M(Br) are the molar masses of carbon and bromine and mt0 is the

mass of the sample measured on the digital balance.

)(

)()11(]/[

BrMm

CkMhhCBr

to

to Equation 5.2

5.1.3. Dimensional Change of Highly Orientated Pyrolytic Graphite

Single crystal measurements were carried out using 12 x 12 x 4mm

HOPG grade ZYH obtained from Momentive Performance. Selected

properties of the graphite are given in Table 5.1. HOPG is not strictly

Page 123: Using Intercalation to Simulate Irradiation Damage of

123

speaking a single crystal, it is in fact a polycrystal made up of very

well aligned crystals. The mosaic spread is a number given to quantify

the alignment of the crystals as shown schematically in Figure 5.6.

Relative to other grades available this is a poor grade of HOPG.

However it was not possible to obtain a sample with such large

dimensions of a better grade. A higher degree of mosaicity means the

properties will slightly deviate from that of a perfect single crystal. It

was decided though, that larger dimensions rather than a better

grade would have a more significant effect on reducing the overall

error in the experiment.

The sample is shown schematically in Figure 5.7 to highlight how axes

are labelled with respect to the sample. The a-axis is parallel with the

graphite layer planes and the c-axis is perpendicular to the layer

planes.

Page 124: Using Intercalation to Simulate Irradiation Damage of

124

Table 5.1 Properties of ZYH HOPG

ZYH grade HOPG[160]

Density 2.255-2.265g/cm3

Spacing of {002} planes 3.355-3.359Å

Thermal Expansion a axis – slightly negative

c axis – 20 x 10-6 (20 - 120°C)

Electrical Resistivity a axis – 3.5 – 4.5x10-5Ωcm

c axis – 0.15 – 0.25 Ωcm

Mosaic Spread 3.5° ±1.5°

Figure 5.6 Schematic of crystal arrangement in HOPG

Figure 5.7 Schematic of HOPG sample with

orientation of axes shown

The results of the HOPG dimensional change experiment are shown

in Figure 5.8 and tabulated in Table 5.2 . The a-axis and c-axis

measurements were recorded on two separate pieces of graphite;

unfortunately it was not possible to measure the two dimensions at

once with the available equipment.

There is a very large expansion in the c-axis with little bromine

absorbed. As discussed in Chapter 2 this is due to bromine creating

interstitial layers in the graphite. The bromine progressively opens

the graphite planes as staging progresses in the structure. The

c

a

4mm

12mm

Page 125: Using Intercalation to Simulate Irradiation Damage of

125

outermost layers become fully brominated first as the boundary of

the sample has lower elastic restraint[96].

There are three curves plotted with the c axis expansion data. The

solid blue line is the theoretical dimensional change curve derived

from the results of Eklund [28], and shown in more detail in [28]. The

dotted red line is a linear best fit curve to all but the first data point.

The black dashed line is a binomial best fit curve including all data

points.

The theoretical curve starts at the origin and ends at the maximum

bromine to carbon ratio C8Br or 12.5%.The dimensional change

induced in a unit cell with the maximum bromine concentration is

given in Equation 5.3, where D.C.Theory is the theoretical dimensional

change, dc is the d spacing between two graphite layers and dCBr is

the spacing between two graphite layers intercalated with bromine.

100..

C

CBrCTheory

d

ddCD

Equation 5.3

The two curves which fit the data describe two different scenarios.

The red linear curve fit implies that the first data point is erroneous.

The gradient of the theoretical expansion rate and the linear curve fit

are very close. The binomial curve has a superior R2 value suggesting

a better fit to the data.

Page 126: Using Intercalation to Simulate Irradiation Damage of

126

y = 16.33x - 5.5978

R2 = 0.9492

y = 0.1961x - 0.2179

y = 16.393x

R2 = 1

y = 3.5784x2 + 6.9335x + 0.1539

R2 = 0.99

-5

0

5

10

15

20

25

30

-2 0 2 4 6 8 10

% Br/C

% S

trai

n

c axis

a axis

c axis theory

Figure 5.8 Dimensional change of bromine intercalated HOPG

Table 5.2 Tabulated Results of HOPG dimensional change

Time

step

h1

(cm) h2 (cm)

Dimensional

Change (%) %[Br/C]

a axis 0 10.897 3.372 0 0

1 10.703 3.18 -0.02676 3.520053

2 10.638 3.005 1.444816 6.728434

3 10.52 2.885 1.471572 8.928467

c axis 0 8.032 7.488 0 0

1 8.009 7.435 11.45631 0.971681

2 8.005 7.415 14.56311 1.338353

3 8.015 7.405 18.4466 1.521689

4 8.02 7.4 20.38835 1.613357

5 8.03 7.395 23.30097 1.705025

6 8.031 7.389 24.66019 1.815027

Page 127: Using Intercalation to Simulate Irradiation Damage of

127

The curve fitting to the data suggests two different possible physical

processes. The linear curve suggests that HOPG is very close to an

ideal crystal expanding within the bounds of error. The binomial

expression puts forward the possibility that there is porosity within

HOPG and this is accommodating some initial expansion. Figure 5.9

shows the microstructure of HOPG and there does appear to be

Mrozowski cracks present.

Table 5.3 shows the steps in calculating a theoretical value for the

mass of bromine required for the measured expansion. By taking the

dimensions of the bromine layer in a stage II bromine intercalation

compound unit cell from the work of Eklund et al[161], it is possible

to extrapolate the mass of a complete mono-layer of bromine atoms

adsorbed to the sample surface. The interlayer spacing induced by an

interstitial layer is 7.05Å[48]. By measuring the expansion of a sample

it is possible to estimate the number of bromine layers intercalated.

Given the theoretical number of bromine layers intercalated and the

mass of a mono-layer it is possible to deduce a theoretical increase in

mass.

The theoretical mass is calculated to be 6.60g which is slightly higher

than the measured value of 6.15g. The percentage difference is

plotted against bromine concentration in Figure 5.10 . The graph

shows that for a given change in mass the measured change in length

is less than the theoretical change in length. In other words there

appears to be some accommodation of expansion which decreases as

Page 128: Using Intercalation to Simulate Irradiation Damage of

128

the bromine concentration increases. This supports the theory that

there are Mrozowski cracks present in HOPG and they do influence

the dimensional change by restricting the measured dimensional

change.

Table 5.3 Analysis of c axis expansion

Original sample height (m) 5.44E-03

Intercalated sample height (m) 6.72E-03

interlayer spacing graphite (m)[160] 6.72E-10

interlayer spacing intercalated bromine (m)[48] 7.05E-10

Number of graphite layers 1.62E+07

Number of bromine layers 3.31E+06

Area containing 4 Br atoms (m

2)[161] 3.84E-19

Area of layer (m2) 1.44E-04

Number of bromine atoms in a layer 1.50E+15

Mass of one layer of bromine atoms (g) 1.99E-07

Theoretical mass increase (g) 6.60E-01

Measured mass increase (g) 6.15E-01

Figure 5.9 HOPG microstructure

Page 129: Using Intercalation to Simulate Irradiation Damage of

129

-60

-50

-40

-30

-20

-10

0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

[Br/C]

% d

iffe

ren

ce

be

twe

en

th

eo

reti

ca

l a

nd

me

as

ure

d

mass

Figure 5.10 Difference between theoretical and measured mass

Figure 5.8 shows that a small change in the a-axis is measured, this is

an unexpected result. It is thought that the most likely explanation

for this is that misaligned crystals are expanding in the a direction.

Given the mosaic spread is 3.5° ±1.5° it is most probable that the

average graphite plane is not parallel to the sample surface but

slightly skewed. The angle created by the expansion vectors as shown

in 5.11 suggests that the true c-axis is 0.8° of the normal of the

sample surface.

Page 130: Using Intercalation to Simulate Irradiation Damage of

130

5.11 In plane expansion

Figure 5.12 shows dimensional change data for irradiated HOPG. The

big difference to note when comparing irradiation data to

bromination data, as shown in Figure 5.8 , is that there is no

shrinkage in the a-axis for intercalated samples. Though it may be

anticipated that there would be a Poisson’s effect shrinking the a-axis

as the c-axis expands this was not measured, therefore any

occurrence of this should be considered to be very small. The

contraction in the a-axis of irradiated specimens is considered to be

due to contraction of graphite planes caused by the creation of holes

in the lattice from displaced atoms[162]. Given the understanding of

intercalation laid out in Chapter 2 it would be unexpected to see this

occur in intercalated samples.

The irradiation data as with the intercalated data has an inflection in

the expansion curves. This suggests that in both instances there are

processes at work which either inhibit early growth or enhance later

growth in the c axis. It can be seen there is also an increase in the

contraction in the a-axis of irradiated specimens as dose increases.

This change in c-axis expansion is normally ascribed to

c expansion

a expansion

Page 131: Using Intercalation to Simulate Irradiation Damage of

131

accommodation of c axis expansion within Mrozowski cracks. This

however fails to explain the increase in the rate of contraction seen

in the a-axis of irradiated samples. It has been suggested that the

contraction in the a-axis is due to larger vacancies being less able to

capture interstitial atoms[36]. As the dose increases there will be an

increase in the number of large vacancies and therefore fewer

instances of holes and interstitials annihilating. Should this be the

case it will also affect the expansion rate in c-axis. Given that the

increase is seen in brominated and irradiated specimens, it seems

reasonable to assume that closure of porosity is at least a

contributing factor.

-15

-10

-5

0

5

10

15

20

25

30

0.0E+00 4.0E+20 8.0E+20 1.2E+21

Dose (EDN)

% S

train

600C in c

600C in a

430C in c

430C in a

Figure 5.12 Irradiation induced dimensional change of HOPG[36]

Page 132: Using Intercalation to Simulate Irradiation Damage of

132

5.1.4. Dimensional Change of Polycrystalline Graphite

The experimental procedure for the polycrystalline experiment is

much the same as for the HOPG experiment. The samples are larger,

75mm long cylinders with a diameter of 10mm and therefore they

are also heavier. This required the use of a different spring as

described earlier in Section 5.1.1.

Three different grades of graphite were used for the polycrystalline

experiment; PGA, Pechiney, and Gilsocarbon. The microstructural

characteristics of these materials are detailed in Chapter 3. Briefly

PGA and Pechiney are two anisotropic graphites with needle like filler

particles while Gilsocarbon is a semi-isotropic graphite with spherical

filler particles.

Figure 5.13 details the results of dimensional change of the extruded

graphites. The extruded graphites have the highest Bacon anisotropy

factor and, as this implies, there is a significant difference in the WG

and AG measurements. The Bacon anisotropy factor gives a ratio

related to the proportion of crystal planes orientated in a particular

direction and as the previous experiment shows, intercalative

dimensional change is driven by c-axis expansion. In the polycrystal

experiments the layer planes are predominantly stacked along the

AG direction and it is this direction which exhibits the largest

dimensional change.

It is possible to measure the anisotropy of PGA using both Bacons

method and by finding the ratio of dimensional change induced by

Page 133: Using Intercalation to Simulate Irradiation Damage of

133

intercalation. The anisotropy due to intercalation is 2.74 ± 0.88, this is

double the Bacon anisotropy factor of 1.44. This suggests that

intercalative expansion is affected by factors other than just plane

orientation. This could be due to microstructural features such as

porosity. There will be a general alignment of porosity with the layer

planes, this will mean that there is a larger accommodation of

expansion between the AG direction than the WG direction which

will affect the anisotropy.

The sample of Pechiney graphite came from a fuel sleeve with

dimensions such that it was only possible to extract a sample aligned

with the grain. The rate of expansion is interesting as it appears to be

too low. The Bacon anisotropy factor for Pechiney, 1.19 is lower than

PGA 1.44. It would therefore be expected that the higher proportion

of misaligned crystallites would cause a higher rate of expansion WG.

Furthermore, Pechiney has a lower porosity volume than PGA.

However the expansion rates are remarkably similar. It is unfortunate

that it was not possible to measure the expansion in the opposing

grain direction to gain a fuller understanding of Pechineys

dimensional change characteristics. This is most likely due to

accommodation of expansion by porosity as will be discussed later.

Page 134: Using Intercalation to Simulate Irradiation Damage of

134

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

0 0.5 1 1.5 2 2.5 3

% Br/C

% S

tra

in

PGA (AG) 1

PGA (AG) 2

PGA (WG) 1

PGA (WG) 2

Pechinay

Figure 5.13 Dimensional change of extruded polycrystalline graphites

Figure 5.14 shows the dimensional change induced in the semi

isotropic graphite Gilsocarbon and compares it with PGA (AG). It is

interesting to note that Gilsocarbon experiences a slightly higher rate

of expansion than the PGA sample cut perpendicular to extrusion

despite PGA having the denser layer plane arrangement. This could

be due to the increased porosity in the PGA samples 27% as opposed

to 19% for Gilsocarbon.

There is a marked increase in the dimensional change rate of the

graphites as the bromine content increases. This trend is also seen

with the dimensional change of HOPG when brominated, this

suggests that a closure of porosity affects both experiments. This is

because as crystallites experience dimensional change there is

initially room to accommodate some of the expansion within

Page 135: Using Intercalation to Simulate Irradiation Damage of

135

microstructural porosity, however as the porosity closes there is less

available space within the microstructure to accommodate the

crystal expansion. This effect will be discussed in more detail in the

following section.

Figure 5.14 Dimensional change of PGA and Gilsocarbon

Figure 5.15 shows the dimensional change induced in irradiated PGA

and Gilsocarbon specimens. The PGA data presented is for AG and

WG samples irradiated at 600°C, the Gilsocarbon data presented has

been irradiated at 430°C and 600°C.

The important difference between intercalation and irradiation

damage is that intercalation damage increases in size whereas

irradiation initially causes a decrease in dimensions. There are two

important reasons for this; first intercalation adds mass to the

sample whereas irradiation damage rearranges the existing mass and

Page 136: Using Intercalation to Simulate Irradiation Damage of

136

second the a-axis contraction of graphite crystals in irradiation

damage is not simulated with intercalation. The a-axis contraction is

a point worth considering because it implies that the stress

generated in the a-axis are considerably stronger than in the c axis.

Despite the larger dimensional changes induced in the c-axis of HOPG

specimens, the a-axis shrinkage is the dominant term. This is because

the c-axis expansion can be accommodated initially within the

Mrozowski cracks however the relief mechanism in the a-axis, creep,

has less influence.

When intercalated, Gilsocarbon expands at a higher rate than PGA

AG samples. This is opposite to what occurs under irradiative

conditions. This suggests that the a-axis shrinkage has a more

significant effect within the Gilsocarbon microstructure. Under pure

expansion conditions Gilsocarbon will expand at a higher rate than

PGA. However, introducing a-axis shrinkage causes PGA to expand

quicker. This could be the result of the spherical filler particles in

Gilsocarbon, a-axis shrinkage will generate hoop stresses which

oppose radial c-axis expansion. PGA filler particles on the other hand

expand relatively freely in the c-axis and shrink in the a-axis.

Page 137: Using Intercalation to Simulate Irradiation Damage of

137

-10

-8

-6

-4

-2

0

2

4

0.00E+00 1.00E+22 2.00E+22 3.00E+22

Dose n cm-2 (EDN)

Dim

en

sio

nal

Ch

ange

(%

)

PGA WG 600C

PGA AG 600C

Gilsocarbon 430C

Gilsocarbon 600C

Figure 5.15 Dimensional change due to irradiation[35]

5.1.5. Analysis of Dimensional Change

As the previous section shows, dimensional change due to

intercalation and irradiation in polycrystalline graphite is a complex

process. There are many influencing factors; the accommodation of

crystal expansion in nanoscale microcracks; the shape and

orientation of filler particles and the distribution of microporosity. By

modelling these factors an insight can be gained into the relevance of

each factor.

The range of open pore volumes in a sample can be measured using

Mercury Porosimetry. Mercury porosimetry is an experimental

technique whereby a sample of graphite is placed in mercury under

Page 138: Using Intercalation to Simulate Irradiation Damage of

138

increasing pressure. As the density increases the mercury penetrates

smaller and smaller pores. By measuring the change in bulk volume

of the mercury, it is possible to gain an insight into the range of

porosity sizes within the graphite microstructure. The plot given in

Figure 5.16 shows that graphite has a large range of pore size from

tiny lenticular nano-cracks to large pores present in the binder and

filler particles [20].

Furthermore Figure 5.16 demonstrates how the range of porosity in a

given grade of graphite can be controlled by manufacturing

processes. Grade AGOT was the graphite used in Chicago Pile 1, it has

a relatively large grain size and impregnation has been carried out.

The effect of this is a large range of porosity sizes including some very

large pores. The grade R-0018 has been baked at high temperature

and under a high pressure, producing a denser graphite with no

measurable pores above 75000 Å. The grade R-0013 is similar to R-

0018 though the grain size is smaller which has the effect of reducing

the average pore size further as well as reducing the proportion of

pores in the 30000 – 75000A range. It has already been suggested in

Section 5.1.3 that pores have an effect on the dimensional change of

HOPG. This section will examine this further.

Page 139: Using Intercalation to Simulate Irradiation Damage of

139

Figure 5.16 Porosity distribution base graphite[20]

To try to understand the role of crystallite orientation and

accommodation on dimensional changes in bromine, a model has

been developed based on the work of Sutton and Howard[163] as

introduced in Chapter 1. Sutton and Howard proposed a model to

relate the crystal coefficient of thermal expansion to the bulk thermal

expansion of PGA graphite. Given in Equation 1.3 and Equation 1.4

they used the Bacon Anisotropy Factor in conjunction with

measurement of crystal CTE to estimate the accommodation porosity

in expansion of the bulk material.

Page 140: Using Intercalation to Simulate Irradiation Damage of

140

acGW CDKCDKCD ...... 21.. Equation 1.3

acGA CDKCDKCD ...... 43.. Equation 1.4

The anisotropy factors used by Bacon are derived from transmission

measurements taken from a single azimuth angle. However as

equipment has improved, gathering texture data with the sample in

an Eulerian cradle has become a standard test, giving more complete

texture data. The transmission measurements used by Bacon gives

the texture data for a single tilt angle whereas the use of a Eulerian

cradle allows data to be collected for a much larger range of tilt

angles. This gives a more complete measurement of the texture in

three dimensions.

The model requires the input of three pieces of information, the

dimensional change of the crystal, the polycrystalline texture data

and the dimensional change of the polycrystalline material. Two

different sources of crystal dimensional change data are used; the

HOPG data given in Section 5.1.3, and using the unit cell dimensions

given by Eklund to derive dimensional change as also discussed in

Section 5.1.3[90]. This is because HOPG data has an intrinsic

accommodation factor. The texture data is given and discussed in

Chapter 4 and the polycrystalline dimensional change data is from

Section 5.1.4.

Equation 5.4 and Equation 5.5 give the theoretical dimensional

change curve for a crystal derived in Section 5.1.3 [28] whilst

Page 141: Using Intercalation to Simulate Irradiation Damage of

141

Equation 5.6 and Equation 5.7 give the measured HOPG dimensional

change curves shown in Figure 5.8. The equations are used to provide

single crystal dimensional change data for the polycrystalline model.

0.. TaCD Equation 5.4

]/[393.16.. CBrCD Tc Equation 5.5

0.. MaCD Equation 5.6

1539.0]/[9335.6]/[5784.3.. 2 CBrCBrCD Mc Equation 5.7

where D.C.x is the dimensional change in the a-axis or c axis and

[Br/C] is the molar ratio of bromine to carbon content in the sample.

As discussed in Section 5.1.3 it is thought that the measured

dimensional change in the a-axis is an artefact of the HOPG sample

microstructure. Therefore the model assumes that D.C.a-HOPG is zero.

The orientation of the stacked layer planes, the c-axis, is determined

by the {002} pole figure of polycrystalline graphite grades. It is

important that the texture data is measured in the correct

orientation, the AG direction should point to zero degrees in the pole

figure.

The pole figure data is defined in polar co-ordinates and describes

the orientation of the crystal axes. The crystal axes need to be related

to a global coordinate system for further modelling. This work uses a

cartesian co-ordinate system for the global co-ordinate system.

Page 142: Using Intercalation to Simulate Irradiation Damage of

142

Figure 5.17 shows the relationship schematically. The angle between

a and b is defined as the tilt angle ψ and the angle between b and c is

defined as the azimuth angle φ. The model also requires the

calculation of θ the angle between a and c.

Figure 5.17 XRD texture orientations

Depending on the polycrystalline orientation to be modelled, WG or

AG, a step function must be applied to the data to ensure the

calculation of theta gives a positive sign. This is because planes

orientated in a direction that would give “negative expansion” would

still have a positive effect on dimensional change. Equation 5.8 gives

the function to model the data for AG and Equation 5.9 gives the

function for modelling WG.

c b

a

Against

Grain Angle of

Measurement

b

d

a

Page 143: Using Intercalation to Simulate Irradiation Damage of

143

360180for ,180

1800for , Equation 5.8

360270for ,270

270180for ,270

18090for ,90

900for ,90

Equation 5.9

Having modified the angles, Equation 5.10 is used to calculate the

angle θ which is the angle between the axis of the measurement and

the axis of interest.

)90()tan(tan

222

1 Equation 5.10

Figure 5.18 and Figure 5.19 show the calculated angle theta for all

instances of XRD intensity measurements. Figure 5.18 shows theta

for AG models, which shows that at azimuth angles 0° and 180° the

crystals are aligned perpendicular to the axis to be modelled. As the

tilt angle decreases all planes regardless of the azimuth angle

become less aligned.

Figure 5.19 shows theta for WG modelling. This shows that when the

azimuth angle is 90° or 270° the crystal planes are orientated

perpendicular to the axis to be modelled, again as the tilt angle

decreases the planes become less aligned.

Page 144: Using Intercalation to Simulate Irradiation Damage of

144

Figure 5.18 Theta derived for against grain calculation

Figure 5.19 Theta derived for with grain calculation

Multiplying the cosine of theta with the proportion of crystal planes

for each theta value gives their contribution to expansion in the axis

of interest. The proportion of crystal planes facing a given angle is

defined by the normalised texture measurement.

Page 145: Using Intercalation to Simulate Irradiation Damage of

145

Figure 5.20 and Figure 5.21 show the effect of applying the PGA pole

figure data given in Chapter 4 to the theta values. This shows how an

anisotropic material such as PGA will react to the two different theta

functions. The against grain plot, Figure 5.20 , has a large proportion

of crystals aligned with low theta angles giving a large potential for

dimensional change. Figure 5.21 on the other hand has a small

proportion of crystals aligned with the low theta angles therefore

giving low dimensional change potential.

Figure 5.20 Effect of crystal orientation on dimensional change against grain

Page 146: Using Intercalation to Simulate Irradiation Damage of

146

Figure 5.21 Effect of crystal orientation on dimensional change with grain

Equation 5.11 predicts the dimensional change of a polycrystalline

material taking the assumption that there are no porosity effects.

Equation 5.12 applies an accommodation porosity factor to give a

more accurate prediction. Alternatively Equation 5.12 can be used

with known results for polycrystalline data to calculate the

accommodation porosity factor.

cos.... mod ICDCD cel Equation 5.11

cos.... .. ICDCD cGW Equation 5.12

Figure 5.22, Figure 5.23 and Figure 5.24 give the change in

accommodation factors with bromine concentration for PGA,

Gilsocarbon and Sutton and Howards thermal expansion results[46].

Sutton and Howards work was carried out on PGA and it can be seen

Page 147: Using Intercalation to Simulate Irradiation Damage of

147

there is rough agreement between the initial HOPG accommodation

factors shown in Figure 5.19, both models are derived from HOPG

measurements. Sutton and Howards accommodation factor appears

to increase at a faster rate pointing to differences in the manner in

which thermal expansion interacts with pores compared to

bromination. Due to the nature of intercalation it is not possible to

calculate the α accommodation factor.

Figure 5.24 shows the accommodation factors Sutton and Howard

measured for PGA [28]. It shows the accommodation factor α to be

more significant than γ for adjusting the thermal expansion model.

This implies that there is a larger contraction in the bulk material

than the shrinkage in the a-axis of the crystal implies. This suggests

that there is a significant Poisson’s effect from the expansion in the c-

axis. There is however a large amount of scatter in the results for α

and it is likely that this is due to large errors arising from the very

small thermal strains which must be measured.

The two methods of deriving the accommodation factor, using that

derived from the unit cell or that derived from HOPG measurements

give different rates of change. Accommodation appears to decrease

with the HOPG derived value whilst it increases for the unit cell

derived values, this is true for both PGA and Gilsocarbon. This is due

to the different crystal expansion rates of the two methods. The unit

cell method says that there will be a linear increase in the crystal c-

axis with intercalated bromine whereas the measured HOPG c-axis

Page 148: Using Intercalation to Simulate Irradiation Damage of

148

expansion value is low initially with an increasing rate of change as

bromine concentration increases.

The measured HOPG accommodation factors have an inherent

accommodation factor which increases (provides less

accommodation) as bromine content increases. The HOPG

accommodation factors for polycrystalline graphites decrease,

provide more accommodation, as the bromine concentration

increases. This suggests either there are more Mrozowski cracks in

polycrystalline graphites, or as bromine concentration increases, the

larger calcination cracks and gas evolution pores start

accommodating expansion or a combination of both.

The accommodation factor for PGA is strongly affected by

orientation. The accommodation factor is around 2.7 times larger in

the AG direction than the WG direction. This is significantly higher

than the bacon anisotropy factor of 1.44. This suggests either

Mrozowski cracks form with a preferred orientation or Mrozowski

cracks are evenly distributed around all crystals but that stresses

generated by the orientation of local crystals affects accommodation.

In other words, large regions of aligned crystals as found in filler

particles are all expanding in the same direction and are only

bounded at the outer edges of the filler particles, so easily fill internal

cracks. The HOPG (AG) curve shows no initial change in value until 0.5

[Br/C].

Page 149: Using Intercalation to Simulate Irradiation Damage of

149

Figure 5.22 PGA bromination accommodation factors

y = 0.0177x2 - 0.1367x + 0.4994

R2 = 0.9995

y = 0.0101x + 0.2367

R2 = 1

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 1 2 3 4[Br/C]

Acc

om

od

atio

n F

acto

r

hopg (AG) accomodation

hopg (WG) accomodation

XRD (AG) accomodation

XRD (WG) accomodation

Figure 5.23 Gilsocarbon bromination accommodation factors

[Br/C]

Page 150: Using Intercalation to Simulate Irradiation Damage of

150

y = 0.7196x + 0.2008

y = 0.2063x + 0.3227

0

0.5

1

1.5

2

2.5

0.1

7

0.1

9

0.2

1

0.2

3

0.2

5

0.2

7

0.2

9

0.3

1

0.3

3

Temperature equivalent [Br/C]

Ac

co

mo

da

tio

n F

ac

tor

gamma block 1

alpha block 1

gamma block 2

alpha block 2

Figure 5.24 PGA thermal expansion accommodation factors[163]

Figure 5.25 gives the predicted dimensional change rates for PGA.

The greatest dimensional change is predicted with the extrapolated

unit cell value. When HOPG data is used, the predicted dimensional

change rate is slightly lower. This is due to the intrinsic level of

accommodation associated with the HOPG data. Applying the

accommodation factor brings the model in line with the data.

Page 151: Using Intercalation to Simulate Irradiation Damage of

151

0

2

4

6

8

10

12

14

16

18

20

0 0.5 1 1.5 2 2.5 3 3.5

[Br/C]

Dim

en

sio

nal

Ch

an

ge (

%)

Measured Data (WG)

Measured Data (AG)

Dimensional Change (WG)

Dimensional Change (AG)

XRD Dimensional Change (WG)

XRD Dimensional Change (AG)

Dimensional Change with Accomodation Factor(WG)

Dimensional Change with Accomodation Factor(AG)

Figure 5.25 PGA predicted dimensional change

Figure 5.26 shows the modelled dimensional change curves for

Gilsocarbon graphite. The curves are shown for the predicted

dimensional change utilising both the unit cell derived factor and the

measured HOPG factor. There is good agreement between the

resulting curves. The anisotropy factor for Gilsocarbon is close to

unity, therefore there is very little difference between the WG and

AG curves, though unfortunately there is insufficient data to obtain a

comparison of the WG and AG accommodation factors.

Page 152: Using Intercalation to Simulate Irradiation Damage of

152

0

2

4

6

8

10

12

14

16

18

20

0 0.5 1 1.5 2 2.5 3 3.5

[Br/C]

Dim

en

sio

na

l C

ha

ng

e (

%)

Gilsocarbon Measured

HOPG Dimensional Change (AG)

HOPG Dimensional Change (WG)

XRD Dimensional Change (AG)

XRD Dimensional Change (WG)

HOPG dimensional change with accomodation factor (AG)

XRD dimensional change with accomodation factor (AG)

Figure 5.26 Modelled dimensional change of Gilsocarbon

As bromine penetrates the graphite microstructure an increasing

amount of porosity is filled by expanding crystals. Figure 5.27 shows

the percentage of total predicted expansion that is accommodated

within polycrystalline microstructures. It shows there is slightly more

accommodation in Gilsocarbon than in PGA (AG). This is in agreement

with the discussion in Section 5.1.5. There is a difference in the

accommodation factors for PGA (WG) and (AG) at higher bromine

concentrations. This suggests that either there is less accommodation

porosity WG than AG or the local microstructure causes the WG

cracks to close up further. The accommodation porosity in the WG

direction comes from misaligned crystals and there are fewer of

these in an anisotropic material. Therefore there will be a larger

restraining force on these layers causing the microcracks to

accommodate more expansion. It is probably a combination of both

Page 153: Using Intercalation to Simulate Irradiation Damage of

153

factors as the microstructural features which would cause the cracks

to close further are probably the same ones that cause the cracks to

exist in the first place.

0

2

4

6

8

10

12

14

0 0.5 1 1.5 2 2.5 3 3.5

[Br/C]

% A

cco

mo

dat

ion

Vo

lum

e

Gilsocarbon

PGA (WG)

PGA (AG)

Figure 5.27 Accommodation Volume

Figure 5.28 , Figure 5.29 , and Figure 5.30 show the effect of using

bromine intercalation accommodation factors to model irradiation

data for low and high doses respectively. The bromination data has

been scaled such that the dimensional change rate of the irradiated

and brominated data are equivalent using Equation 5.13.

]/[..

]/[

..

CBr

d

CdD

CBrd

CdD

DosecIr

cBr

Equation 5.13

Figure 5.28 shows predicted dimensional change at a low irradiation

temperature of 200°C using HOPG data given in [28]. The irradiation

data used to compare the predicted dimensional change with the

Page 154: Using Intercalation to Simulate Irradiation Damage of

154

actual dimensional change is collected in the DIDO and PLUTO

reactors [28]. The predicted curve is slightly higher than the actual

data and this can be attributed to two factors. From the difference in

temperature of the predicted and measured data it is anticipated

that the predicted data will be slightly higher as shown in Figure 1.16

[28]. Furthermore the effect of shrinkage in the a-axis cannot be

accounted for with this method. Nevertheless there is good

agreement between the model and experimental data.

Figure 5.28 Prediction of Irradiated Dimensional Change using intercalation accommodation factors;

simulation of low neutron dose[28]

At higher irradiation temperatures the model does not work so well.

At low doses there is a small amount of expansion predicted using

the bromination accommodation factors. The values are not

unexpected as it is not possible to model the a-axis accommodation

factor and it seems reasonable to assume that these two factors

Page 155: Using Intercalation to Simulate Irradiation Damage of

155

combine to result in the small dimensional change that is seen.

However when the model is extrapolated out as shown in Figure 5.30

it is of little use. This is because of a fundamental difference in the

two processes. When irradiated to a dose of 1.37x1021 EDN at 430°C

a c-axis expansion of 24% is induced however the volume of the

sample reduces by around 5%. Compare this to bromination data

where a [Br/C] ratio of 1.81 also induces 24% expansion in the c axis

and a corresponding 24% increase in volume.

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

-1E+20 1E+20 3E+20 5E+20 7E+20 9E+20 1.1E+21 1.3E+21 1.5E+21

[Br/C] normalised to dose EDN (n cm-2)

Dim

en

sio

nal

Ch

ange

(%

)

Dimensional Change Gilsocarbon @ 430C XRDBaseDimensional Change Gilsocarbon @ 600C XRDBaseDimensional Change 430C Irradiation

Dimensional Change 600C Irradiation

Dimensional Change Gilsocarbon @ 430C HOPGBaseDimensional Change Gilsocarbon @ 600C HOPGBase

Figure 5.29 Prediction of Irradiated Dimensional Change using intercalation accommodation factors;

simulation of low neutron dose

Page 156: Using Intercalation to Simulate Irradiation Damage of

156

-5.00

-3.00

-1.00

1.00

3.00

5.00

7.00

9.00

0 5E+21 1E+22 1.5E+22 2E+22 2.5E+22 3E+22

[Br/C] normalised to dose EDN (n cm-2)

Dim

en

sio

nal

Ch

ange

(%

)

Dimensional Change Gilsocarbon @ 430C XRDBaseDimensional Change Gilsocarbon @ 600C XRDBaseDimensional Change 430C Irradiation

Dimensional Change 600C Irradiation

Dimensional Change Gilsocarbon @ 430C HOPGBaseDimensional Change Gilsocarbon @ 600C HOPGBase

Figure 5.30 Prediction of irradiated dimensional change using intercalation accommodation factors;

simulation of high neutron dose

5.2. Microstructural Experiment

The last two sections have shown that there are key parallels

between bromination and irradiation damage. The most important

parallel being the large disparity in single crystal and polycrystalline

volume changes due to porosity in accommodating expansion and

the orientation of crystals. To investigate the behaviour further, it

was decided to observe bromination in real time using XCT to

observe the microstructural strains due to the crystallite growth. To

the authors knowledge this had not been done before.

A bromination rig was designed for use in a high resolution

tomography experiment; the setup is shown in Figure 5.31 .

Synchrotron tomography was used primarily because the fast data

Page 157: Using Intercalation to Simulate Irradiation Damage of

157

acquisition rate meant that the small amount of expansion which was

expected to occur during each tomographic scan, around eight

minutes, would be below that of the resolution of the camera

allowing high quality reconstructions.

The samples used were 1 x 1.5 mm. This allowed a resolution of

0.7µm, which was a compromise between obtaining the highest

resolution possible and machining constraints. The synchrotron was

setup to have a beam power of 19.5KeV. The camera collected each

image in 325ms and there were 1501 projections for each data set

meaning each data set was collected in just over eight minutes.

Once the bromination rig was prepared, it was loaded onto the

sample stage. At this point there is no bromine in the sample

chamber and a scan is performed on each sample. The stop cock is

then opened releasing bromine into the sample chamber. A scan is

then performed every 30 minutes on each sample.

The data is reconstructed using in house algorithms. The machine

was setup to use the Modified Bronnikov Algorithm (MBA) for

reconstruction [164]. However, as the bromine intercalated the

sample the MBA technique became unsuitable. The MBA algorithm is

optimised for samples with little absorption contrast, virgin graphite

being a prime example. However, as bromine enters the structure

there are large differences in absorption around the sample.

Therefore, the samples were reconstructed with the standard

absorption contrast reconstruction algorithm.

Page 158: Using Intercalation to Simulate Irradiation Damage of

158

Figure 5.31 Bromination rig in position. Highlighted are a) X-ray source b) shutter c) rig and d) camera

5.2.1. Two dimensional Analysis

Digital Image correlation was used to analyse the two-dimensional

radiographs in order to estimate the bulk dimensional change

experienced by the samples, a technique which has been used

previously in studies of rocks[165]. This was achieved using DaVis V6

from LaVision software. An example image is given in Figure 5.32 .

The overall growth was defined as the difference in expansion of the

average displacement vector at Z=200 and at Z=900. This calculation

was performed on 100 out of 1501 of the radiographs and taken as

an average.

d c b a

Page 159: Using Intercalation to Simulate Irradiation Damage of

159

Figure 5.32 Digital image correlation performed on PGA radiographs

The results are given in Figure 5.33 , Figure 5.34 and Figure 5.35 for

PGA, Gilsocarbon and Pechiney respectively. As with the bulk

expansion experiments described in Section 5.1.4, there is an overall

general expansion. PGA achieved the largest dimensional change of

4% followed by Pechiney 0.4% and finally Gilsocarbon 0.3%. This

shows that when all samples are in the same chamber in the same

concentration of bromine, there is a difference in dimensional

change rates. Primarily this is because the larger volume of open

porosity in the graphite makes it easier for bromine to permeate the

microstructure. Further to this is the fact that PGA has larger crystals

with better crystallinity which means that once the bromine has

Page 160: Using Intercalation to Simulate Irradiation Damage of

160

reached a crystal for intercalation it can permeate the crystal with

less difficulty[106].

In all cases there was an initial shrinkage, this may possibly be

attributed to the endothermic intercalation reaction[28]. However,

the cooling required for the contraction observed in these

experiments would need to be very large. Therefore, the author

considers that the most likely explanation is an error introduced by

the experimental technique. An initial scan was taken with no

bromine in the sample chamber then a stopcock was released for the

bromine to enter the chamber and this may have caused a significant

movement that would be recorded by the scans. If the movement

sample moves away from the x-ray source between the two scans

the sample will appear smaller and the digital image correlation

software will record an apparent contraction. Furthermore any

rotational displacements will distort the radiograph and therefore

provide erroneous results, this can be overcome by reconstructing

the data and carrying out 3D DIC.

The PGA and Pechiney samples in Figure 5.33 and Figure 5.34 appear

to have a bilinear dimensional change curve. Though it can’t be

directly correlated to the earlier measurements as it is not possible to

know the [Br/C] of the sample, it does appear to follow a similar

trend to the AG bulk samples. The bulk dimensional change is of a

similar magnitude for PGA in the microstructural experimental rig

and the polycrystal experimental rig.

Page 161: Using Intercalation to Simulate Irradiation Damage of

161

The dimensional change curve for Gilsocarbon is a lot more

complicated. There appears to be significant noise in the data,

though the general trend appears to be more linear than the

anisotropic graphites. The noise was caused by out of plane

movement between the images causing errors in the image analysis.

-1

0

1

2

3

4

5

0 100 200 300 400 500 600

Time (mins)

Dim

en

sio

nal ch

an

ge (

%)

Figure 5.33 Dimensional change calculated from radiographs for PGA

Page 162: Using Intercalation to Simulate Irradiation Damage of

162

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0 100 200 300 400 500 600 700

Time (mins)

Dim

en

sio

na

l c

ha

ng

e (

%)

Figure 5.34 Dimensional change calculated from radiographs for Gilsocarbon

-3.00E-01

-2.00E-01

-1.00E-01

0.00E+00

1.00E-01

2.00E-01

3.00E-01

4.00E-01

5.00E-01

0.00E+00 1.00E+02 2.00E+02 3.00E+02 4.00E+02 5.00E+02 6.00E+02 7.00E+02

Time (mins)

Dim

en

sio

nal ch

an

ge (

%)

Figure 5.35 Dimensional Change calculated by digital image correlation for Pechinay

Page 163: Using Intercalation to Simulate Irradiation Damage of

163

5.2.2 Three Dimensional Analysis

The PGA sample was selected for further microstructural analysis as

the sample underwent similar dimensional change to the

polycrystalline experiments. This was performed using digital volume

correlation, a technique which can take into account out of plane

displacements and remove these errors from the analysis.

Before digital image correlation could be performed, some

preprocessing of the images was required. The experimental rig was

not perfect and the sample would have displacements associated

with it that weren’t due to intercalation. These rotations were

removed using the image analysis software ImageJ.

The reconstructed image stack was resliced to obtain the image stack

as slices in the XZ plane. The images were rotated so that the sample

always remained vertical. The new Image stack was resliced to obtain

images in the YZ plane. These images were also rotated to keep the

sample vertical in this orientation as well. The images were then

resliced and into the XY plane the image stacks were cropped so that

all image stacks were the same size.

Unfortunately each of the rotation step reduces the image quality.

This is because ImageJ interpolates the new greyscale value by

rotating the old pixels and calculating the greyscale value[166]. To

reduce the effect of interpolation artifacts a bicubic filter was used.

Page 164: Using Intercalation to Simulate Irradiation Damage of

164

A bicubic filter is a method which calculates the value of a pixel in a

modified image by interpolating the value from the surrounding

pixels. Figure 5.36 describes the process in 1 dimension. By fitting a

curve to the greyscale values of the nearest neighbour pixels the

algorithm can calculate a value for the interpolated pixel value

highlighted by the red dotted line. For image analysis this is method

is applied in two dimensions and the interpolated value calculated as

a point on a surface[167].

1 2 3 4

Pixel

Figure 5.36 Bicubic filter

Figure 5.37 details a cross section through the centre of the PGA

sample. The majority of the sample is comprised of binder matrix.

This is characterised as a region of disordered crystallites permeated

with gas dendritic evolution pores. In the top right of the sample

there is a filler particle which has aligned crystallites and lenticular

calcination cracks. The sample area highlighted is 0.9 x 0.6mm. The

Page 165: Using Intercalation to Simulate Irradiation Damage of

165

colour of the vectors highlights the magnitude of its component into

the page. Green signifies the vector is pointing into the page 1μm and

yellow indicates 1μm out of the page.

Figure 5.37 Microstructure of PGA tomography sample, before bromination

The resulting series of images are given in Figure 5.38 . The

progression of bromine through the microstructure can be observed

through the change in grey scale of the images. In the early images

the bromine first comes into contact with the graphite but no

significant intercalation has taken place.

Calcination crack

Binder

Matrix

Gas Evolution Pores

Filler

Particle

Page 166: Using Intercalation to Simulate Irradiation Damage of

166

After two hundred and sixty minutes, there is the first sign of

intercalation in the binder matrix, highlighted by A, but it is not until

four hundred and forty minutes that there is any significant

bromination of the filler particle, highlighted by C. At this point the

binder matrix is nearly fully brominated.

The binder is clearly brominating before the filler particle. Though the

filler particle is more ordered than the binder, there is a higher partial

pressure of bromine within the binder matrix due to the pervading

open pore network within the binder. This brings to light a problem

in using bromine intercalation to simulate irradiation damage. For a

given a uniform dose profile irradiation will damage all crystals

equally. Bromine intercalation preferentially damages crystals that

are next to pores and are easy to access.

An interesting observation is that the displacement vectors also show

that the large amount of bromine intercalated causes a low level of

expansion. The binder matrix is responsible for the early expansion of

the material. On the other hand, the small amount of bromine which

penetrates the filler particles later in the series causes a large amount

of dimensional change. This implies that the aligned crystals in the

filler particles are the main driver in bulk dimensional change,

whereas intercalation of the disordered crystals has little overall

effect.

A further interesting point to note is the pore highlighted in B. This is

a calcination crack, a long lenticular crack running parallel to the WG

Page 167: Using Intercalation to Simulate Irradiation Damage of

167

direction. As bromine intercalation occurs the pore closes, this

indicates that accommodation of crystal expansion is not limited to

Mrozowski cracks but in the case of bromine intercalation also occurs

in the larger microporosity.

Page 168: Using Intercalation to Simulate Irradiation Damage of

168

a) t=40 mins

First bromine can be seen adsorbing at pore edges

b) t=40 mins

c) t = 100 mins d) t = 140mins

B B

B B

A A

A A

Page 169: Using Intercalation to Simulate Irradiation Damage of

169

e) t=200 mins f) t =260 mins First signs of bromine penetrating binder

matrix

g) t = 320 mins Binder is nearly fully brominated by contrast

there is still little bromine in the filler

h) t = 380 mins

B

B B

B

A A

A A

Page 170: Using Intercalation to Simulate Irradiation Damage of

170

i) t=440 mins First signs of significant dimensional change

in filler particle highlighted in red circle

j) t=500 mins

k) t=560 mins

Figure 5.38 Digital Volume Correlation of PGA Bromination

B B

B

A A

A

C C

C

Page 171: Using Intercalation to Simulate Irradiation Damage of

171

Because bromine has a higher x-ray attenuation coefficient than

graphite, regions intercalated with bromine become lighter. Figure

5.39 plots the greyscale distribution for each time step. There are

three distinct peaks on the graph; the lowest value signifies the

region with the lowest absorption which is porosity; the second peak

signifies regions of virgin graphite; the highest peak on the greyscale

axis signifies regions of graphite intercalated with bromine. It can be

seen that as the time increments increase there is an increase in the

magnitude of the bromine graphite peak and a decrease in the pore

peak. This could be due to graphite intercalated with bromine

closing porosity as well as bromine condensing within the pores.

-10000

0

10000

20000

30000

40000

50000

60000

70000

80000

0 10000 20000 30000 40000 50000 60000 70000

Grey Scale Value

Co

un

ts

t10

t40

t70

t100

t140

t200

t260

t320

t380

t440

t500

t560

Figure 5.39 Change in grayscale of brominated samples

Pore

Graphite

Bromine in

Graphite

Page 172: Using Intercalation to Simulate Irradiation Damage of

172

5.3. Conclusions

This chapter has investigated dimensional change due to

bromination, this is important as it is a very useful way of

introducing interstitial defects, which in turn induce

microstructural defects, in a laboratory setting. The results

presented in the chapter show that the dimensional change of a

bulk material is a complex process affected by many different

factors.

The work shows that bromine intercalation of single crystal

graphite produces significantly larger strains than seen in

polycrystalline graphites. This is primarily due to porosity within

graphite. Though the graphite in the binder phase is less ordered

than filler phase it is still of a high enough graphitic quality that

intercalation occurs. Therefore, the surrounding partial pressure

of bromine is the dominant factor in determining which regions of

graphite intercalate first. The microstructure of graphite is also

shown to have a significant effect on the dimensional change rate.

This work suggests that porosity, crystal orientation, and average

lattice spacing have an influence on the rate of expansion. These

microstructural factors have been combined to create a model

which gives good agreement at low irradiation temperatures,

however, at higher irradiation temperatures the model gives poor

agreement with experimental data and it is thought this is

primarily to do with a lack of information about a-axis

deformations gained from intercalation experiments.

Page 173: Using Intercalation to Simulate Irradiation Damage of

173

A tomographic experiment has been carried out to investigate the

development of microstructural strains. The important result here

is that dimensional change is driven by interstitial damage of the

filler particles whilst the damage to the binder matrix has a rather

small affect on the overall dimensional change. There exists one

key difference between irradiation damage and intercalation

damage and which is that intercalation is a damage process which

progresses through the microstructure.

The work presented here has only scratched the surface of what is

possible with using bromination as a simulation technique. It is

suggested that future studies should try to determine the layer

spacing in the binder matrix and the filler matrix and how these

two parameters develop with bromination. This could provide key

information regarding the potential of graphite to accommodate

damage to the microstructure. Further development of techniques

such as x-ray diffraction contrast tomography would allow better

models to be built which could take into account local crystal

orientations which could be applied to three dimensional finite

element models and provide an insight into how different regions

interact as strains develop.

Page 174: Using Intercalation to Simulate Irradiation Damage of

174

6. Young’s Modulus

When graphite is placed in an irradiating environment the Young’s

modulus will undergo a series of changes. It is generally considered

that there are two main processes contributing to the change in

modulus as given in Equation 6.1[42].

),()(0

TSTPE

Ed

Equation 6.1

The P(T) term describes “pinning” a process which causes a sharp

initial increase in modulus. The increase in modulus is induced at a

low dose. The increase can be around 50% and this has been

attributed to a change in the C44 crystal shear modulus[44].

The second term S(δ,T) is a structure factor which is influenced by

dose and temperature. The term is describing the effect of crystal

volume changes on the bulk Young’s modulus. The structure term

becomes significant at higher doses. Initially the term describes an

increase in modulus. It is thought that this is due to crystal expansion

closing porosity causing a “tightening” of the structure[57]. As dose

increases further the term describes a decrease in modulus.

Eventually at high enough dose the volume change of crystals will be

such that new porosity starts opening causing the modulus to

decrease.

Page 175: Using Intercalation to Simulate Irradiation Damage of

175

As shown in Chapter 5, crystal expansion can be driven by

intercalation as well as irradiation. This chapter describes a series of

experiments designed to investigate the structure term and

investigate how intercalation will influence this term.

Modulus measurements of nuclear graphite are normally carried out

using dynamic testing rather than static testing. The two methods

give the same results if the static modulus is measured for low strains

0.01 – 0.05%[53, 167]. However, above this the results diverge.

Dynamic modulus measurements are normally taken using piezo

transducers as specified in the American Society for Testing and

Materials (ASTM) standard for the measurement of the graphite

modulus[134]. The work described in this thesis modified the

recommended method by using laser techniques for ultrasound

generation and measurement. This was necessary because of the

harsh environmental conditions of the bromination experiment.

This is not the first work to use laser impact excitation with graphite;

previous work used laser impact excitation in conjunction with

Electro Magnetic Acoustic Transducers (EMAT)[135] to investigate

the effect of thermal oxidation on the Young’s modulus of graphite.

An EMAT is a type of non-contact sensor that measures the

interaction of magnetic fields in a sample and the sensor. This

requires the sensor to be quite close to the sample edge.

Page 176: Using Intercalation to Simulate Irradiation Damage of

176

Laser vibrometry is another non-contact method that can also be

used to measure surface vibrations. Vibrometry has the advantage

that it can be used relatively far away from the sample. A further

advantage particularly relevant for bromination of graphite is that

the sample dimensions can change and the vibrometer can continue

to measure a signal. There is a caveat that the sample surface

remains within the depth of field of the vibrometer, therefore large

dimensional changes require periodic refocusing of the vibrometer.

The chapter details the development of the modified dynamic

young’s modulus testing technique. The modified technique is

applied to measuring changes in brominated graphite. The overall

aim of this is to investigate how the Young’s modulus structure factor

is modified by bromine intercalation.

6.1. Measuring optimal input energy for Young’s modulus measurements by

laser impact excitation

When using lasers to induce an ultrasonic pulse, it is important to

establish the optimal input energy. At low energies the ultrasonic

wave is generated thermoelastically. However, at higher energies an

ablatic waveform is superimposed[168]. Ablation occurs when a high

energy density beam interacts with a material and causes some

material to vaporize[136]. Of the two waveforms, the lower energy

thermoelastic wave induces an amplitude proportional to the input

energy. The higher energy ablatic waveform produces a significantly

larger amplitude in the waveform but it also degrades the sample

Page 177: Using Intercalation to Simulate Irradiation Damage of

177

surface. It is therefore preferable to carry out the experiment at the

edge of ablation regime.

To determine the optimal energy, a simple experiment was carried

out using a 50mm long PGA sample cut in the WG direction, Figure

6.1. The sample was placed on a lab jack with the impact laser and

the vibrometer focused on the opposing 20 x 5mm end surfaces. Care

was taken to ensure the surfaces remained perpendicular to the laser

beams and the two beams remained parallel. An average of five

measurements were taken for each output energy available from the

“Big Sky Nd:YAG laser”, which is 0 to 20mJ at 0.5mJ increments.

Figure 6.1 Ultrasonic pulse measured with laser vibrometer

Figure 6.2 shows the trace recorded from the laser vibrometer

measuring the surface displacement on a graphite sample which has

been struck by the Nd:YAG laser on the opposing surface. Figure 6.3

shows how increasing the input energy affects the signal recorded

with the vibrometer. To understand this graph it is best to first look

at, the captured waveform from the oscilloscope. The x axis is time

and the y axis is the output of the vibrometer measured in volts.

WG sample AG sample

50mm

5mm

20mm

Page 178: Using Intercalation to Simulate Irradiation Damage of

178

Figure 6.3 shows what happens as the energy of the impact laser

increases. Here the x axis is the impact laser energy, time is now the y

axis and the colour map defines the wave amplitude (the y axis in

Figure 6.2). The plot shows a high energy input laser pulse induces an

ultrasonic wave with large amplitude. Though this is easy to measure

it has the undesirable effect of ablation. Therefore an energy level

should be set such that a clear signal is induced with the lowest input

energy. The optimal energy used was chosen as 7mJ.

Whilst carrying out this experiment at higher energies it was

necessary to adjust the sample position to ensure the laser beam was

not hitting a surface previously damaged by ablation. The sample was

moved perpendicular to the laser beams to achieve this. When

moving the sample it was imperative to ensure that the sample

remained perpendicular to the impact laser and vibrometer.

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

-1.00E-05 0.00E+00 1.00E-05 2.00E-05 3.00E-05 4.00E-05

Time (s)

Am

plitu

de (

V)

Figure 6.2 Ultrasonic pulse measured with laser vibrometer

Page 179: Using Intercalation to Simulate Irradiation Damage of

179

Figure 6.3 Effect of Impact Energy on Measured Waveform

6.1.1. Verification of Laser Impact Excitation as a suitable modification of

ASTM C769

To verify the suitability of laser measurements for the determination

of Young’s modulus of graphite, a direct comparison was made

between the standard technique and the laser technique. Two grades

of graphite were used, Gilsocarbon and PGA cut perpendicular to

extrusion. The samples were dimensioned 75 x 5 x 20mm. The end

faces of the samples were polished to produce a smooth reflective

surface using grade 4000 SiC polishing paper. This improves the

transmission of ultrasonic waves between the piezo transducers

whilst also improving the quality of the reflected signal for

vibrometry measurements.

Osc

illo

sco

pe

amp

litu

de

of

ult

raso

nic

wav

e (V

)

Energy of impact laser (mJ)

Page 180: Using Intercalation to Simulate Irradiation Damage of

180

6.1.1.1. Standard Experimental Technique

The first modulus experiments were taken using the standard setup

shown in Figure 6.4. The equipment comprises 5MHz Olympus

transducers connected to an Olympus pulse receiver. The signal is

read by a National Instruments PCI 5124 digitizer card which samples

at 200Ms/s. Averages of three measurements were taken for each

sample. The density was calculated by measuring the sample

dimensions to three decimal places and taking a mass measurement

of four significant figures.

Figure 6.4 ASTM C769 Experimental setup for piezo transducer measurement of Young's modulus

The sample was held between the two transducers using a clamp

along its longest length. A propriety gel, (Sonagel W) was used as a

couplant. The impact wave form and the transmitted wave form are

recorded by the computer as shown in Figure 6.5.

Page 181: Using Intercalation to Simulate Irradiation Damage of

181

The excitation pulse is convoluted with the receiving pulse. The time

of flight is measured as the time between the impact pulse reaching

10% of its maximum amplitude and the receiving pulse reaching 10%

of its maximum amplitude.

Figure 6.5 Time of flight data for modulus measurement by piezo transducer

6.1.1.2. Experimental Setup for Laser Technique

Sonic velocity measurements using the laser technique were carried

out using the experimental setup shown in Figure 6.6 and Figure 6.7.

The ultrasound was generated by focusing a 7mJ laser pulse fired for

50ns at full width half maximum from a Nd:YAG source onto the

graphite sample.

A Polytec vibrometer was mounted on a sliding plate and focused on

to the opposing sample surface. Course focusing of the vibrometer is

carried out using the sliding plate. Finer focusing was performed

using the focus on the vibrometer. The vibrometer has the ability to

Page 182: Using Intercalation to Simulate Irradiation Damage of

182

be setup to measure the displacement or the velocity of a sample

surface. In this case it was set up for changes in surface velocity as

this method is sensitive to small surface displacements[28].

The Q Switch Sync and vibrometer signals were connected to a

500MHz LeCroy digital oscilloscope. The Q-switch is an electo-optical

device that fires the laser, the Q switch sync gives the zero time for

the experiment.

The signal used for time of flight measurement was taken as an

average of 10 pulses. The signal was smoothed further using the

oscilloscopes inbuilt Gaussian filter. The time of flight was recorded

as the time between the 10% height of the first pulse seen in the Q

switch signal and the vibrometer signal. The recorded results are an

average of three such measurements.

Figure 6.6 Experimental setup for measurement of Young's Modulus by Laser Impact Excitation and Laser

vibrometry

Page 183: Using Intercalation to Simulate Irradiation Damage of

183

Figure 6.7 Schematic of experimental setup for measurement of Young's Modulus by Laser Impact Excitation

and Laser vibrometry

6.1.1.3. Results

The results from the two experiments are shown in Figure 6.8. The

results show general agreement though there is a small systematic

difference between the two techniques. The laser technique

produces results which are on average 0.6μs slower with a standard

deviation of 0.54 μs. This may be attributed to the delay in the

electronics of the measurement circuit. The piezo technique

accounts for this by measuring the time delay when the transducers

are brought into contact. A similar process is not possible using

lasers. When using lasers without a sample there is no ultrasonic

wave generated nor is there a signal to record for an arrival time.

The scatter in the time of flight data has been observed by previous

Sample Nd:Yag

Vibrometer

LDD

Labjac

Oscilliscope

PC

Sample Stop

Page 184: Using Intercalation to Simulate Irradiation Damage of

184

authors[169]. They attribute it to microstructural effects in the

graphite.

Figure 6.8 Comparison of sonic velocity measurements by Laser and Piezo Techniques

6.2. Change in modulus of brominated graphite

The dynamic modulus of a material is determined by three

properties; density, Poisson’s ratio and the sonic velocity of a pulse

through the material as derived by Timoshenko[132] and given in

Equation 6.2 [134] .

)1(

)21)(1(2

vEDYM

Equation 6.2

where ρ is the sample density, v is the sonic velocity and here η is

Poisson’s ratio. The work in this thesis has focused on the change in

Page 185: Using Intercalation to Simulate Irradiation Damage of

185

density and sonic velocity. Though the change in Poisson’s ratio has

not been measured or calculated for graphite intercalated with

bromine, it was deemed unnecessary to measure this property for

this work. This decision was based on work on lithium intercalation

which showed the change in Poisson’s ratio to be small[114]. It was

therefore decided that Poisson’s ratio was the least important of the

properties to focus upon.

The following section will present the results of experimental work

carried out to determine the change in density and sonic velocity of

brominated graphite.

6.2.1. Change in density of brominated graphite

The changes in density of brominated graphites were derived from

the measurements carried out in Section 5.1. Gilsocarbon is a semi-

isotropic material[170]. The decision was taken that the dimensional

change in one axis would be the same in the other axes. This gives

Equation 6.3 which describes the change in density of an isotropic

material given the dimensional change in one axis.

2.. xCrDl

m

Equation 6.3

Where m is the mass of the sample, l is the length of the sample for a

given level of bromination, r is the original radius of the sample and

D.C.x is the dimensional change of the sample measured at that level

of bromination.

Page 186: Using Intercalation to Simulate Irradiation Damage of

186

The method used to measure the change in density of PGA is slightly

more complicated given its high anisotropy ratio. To do this a curve

was fitted to the dimensional change of PGA samples given in Section

5.1 giving an equation for the expansion of the sample perpendicular

to the measured direction for a given bromine concentration. This

was applied to the formula given in Equation 6.4.

2)..(

CrDl

m

Equation 6.4

All the symbols here are the same as for equation 5.1 apart from the

strain component..CD .

..CD describes the dimensional change in

the axis perpendicular to the axis from which the length (l) and mass

(m) data is taken.

To calculate the volume change of HOPG, it was assumed that there

was no dimensional change perpendicular to the planes and that all

volume change came from expansion parallel to the graphitic planes.

The justifications for these assumptions are discussed in Chapter 4.

Figure 6.9 shows that when bromine is intercalated with HOPG there

is a rapid increase in density. The density settles around 2.21g/cm3,

just under the ideal single crystal value of 2.25 g/cm3, the ideal single

crystal value. This suggests that crack closure occurs up to around

15% strain, this corresponds to a [Br/C] ratio of 1.4%. After this there

appears to be little change in density which suggests there is little

subsequent change in crack volume.

Page 187: Using Intercalation to Simulate Irradiation Damage of

187

Figure 6.10 shows the change in density of polycrystalline graphites.

At low intercalation levels density increases up to 0.5% to 1.5% strain

depending on the graphite grade. This corresponds to a [Br/C] ratio

of approximately 1%. It is probable that Mrozowski cracks close in

polycrystalline graphites earlier than HOPG due to restraining forces

of the surrounding microstructure. It is therefore assumed that the

initial densification is caused by closure of Mrozowski cracks.

As intercalation strain increases further, the density decreases. This is

against a backdrop of increasing single crystal density, decreases in

the accommodation factors (which imply a closure of porosity) and

tomographic scans showing closure of porosity. The reason for this is

that dimensional change is driven by filler particle expansion. The

samples used for the tomographic scans showed no crack generation

because they were too small and so there was nothing restricting

filler particle expansion.

Figure 6.11 shows a laser confocal micrograph of a debrominated

PGA sample. The filler particle is highlighted in image a). The filler

particle expansion generates a crack which propagates though the

binder along the with grain direction. The generation of porosity in

the binder causes the decrease in bulk density.

Page 188: Using Intercalation to Simulate Irradiation Damage of

188

1500

1600

1700

1800

1900

2000

2100

2200

2300

0 5 10 15 20 25 30 35

% Longitudinal Strain

De

ns

ity

(k

g/m

^3)

Figure 6.9 Change in density of HOPG

1500

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

% Longitudinal Strain

Den

sit

y (

kg

/m^

3)

Gilsocarbon

PGA (AG)

PGA (WG)

HOPG

Poly. (PGA(AG))

Figure 6.10 Density change of polycrystalline graphites

Page 189: Using Intercalation to Simulate Irradiation Damage of

189

Figure 6.11 Laser confocal micrograph of debrominated PGA a) green scale bar 100µm area of image b

highlighted in green b) scale bar 40µm filler particle outlined in green

By fitting curves to the calculated density data curves, equations can

be found which relate density to dimensional change. This is of use

for measurement of Young’s modulus later on. The equations for

change of density are given below.

PGA (AG) )0118.1..0353.0.0133.0( 2

0 AGAG CDCD

Equation 6.5

PGA (WG) )0648.1..0135.0.0046.0( 2

0 WGWG CDCD Equation 6.6

Gilsocarbon )0137.1..0081.0.0019.0( 2

0 CDCD Equation 6.7

6.3. Modulus changes of brominated Graphite

6.3.1. HOPG

To understand modulus changes in polycrystalline graphites subject

to bromination, it is first important to understand the modulus

changes in single crystal graphite. The modulus of single crystals are

defined using the fourth order tensor Cαβστ. The can be defined by

WG

AG

Page 190: Using Intercalation to Simulate Irradiation Damage of

190

stress – strain relations as show in Equation 6.8 or equations of

motion as given in Equation 6.9.

uCT Equation 6.8

)(2 Cv a Equation 6.9

In Equation 6.8 Tαβ describes the stress vector orientated in direction

αβ, uστ is the strain vector in direction στ. In Equation 6.9 the v is the

velocity of a plane wave with polarization φ of the three types of

waves propagated in direction of the unit vector ξ[55].

Measurements were carried out using longitudinal and shear

transducers as shown in Figure 6.4. The setup is carried out according

to ASTM C769[134] and was carried out on HOPG before and after

bromination. The brominated sample was kept separate from the

transducers using 15mm cubed PTFE blocks as shown in Figure 6.12.

Figure 6.12 Piezo setup for brominated samples.

15mm

Sample

PTFE PTFE

Piezo

Transducer

Piezo

Transducer

Page 191: Using Intercalation to Simulate Irradiation Damage of

191

The work was carried out on ZYH grade HOPG purchased from SGI

through Momentive Performance Materials. This is the lowest grade

available which has a mosaic spread of 3.5° ± 0.5°. This made it

affordable to buy a relatively large sample of 12x12x4mm HOPG.

The results are shown in Table 6.4 and Figure 6.13. When comparing

the virgin results with literature values, there is good agreement

between the values for shear modulus C44. However, the values for

C11 and C33 are slightly low[171]. This is most likely due to the grade

of HOPG used. As earlier experiments show, this grade of HOPG has a

relatively high internal porosity which will lower the modulus. As

density calculations in section 6.2.1 show there is porosity within the

graphite which will have the effect of lowering the modulus from the

ideal value.

The calculation for the brominated modulus is carried out by

measuring the time of flight through the PTFE rods and the graphite

sample. The time of flight through the PTFE is two orders of

magnitude larger than the time of flight through the graphite sample.

Despite every effort to do the experiment quickly, a further problem

arises due to the diffusion of bromine out of the sample as the

measurements were taken. This is a rapid process once the external

partial pressure of bromine is reduced by taking the sample out of

the bromination chamber. This is particularly noticeable in the

difference between C11 and C22 measurements which are expected to

be similar, as they are for virgin measurements.

Page 192: Using Intercalation to Simulate Irradiation Damage of

192

Overall it can be seen that there is an increase in modulus associated

with the intercalation of bromine in HOPG. This is most noticeable in

the C11, C22 & C33 modulus which could possibly be explained by

closure of Mrozowski cracks which will transfer a compression wave

faster.

The shear modulus remains relatively unchanged. This implies that

pinning is not a dominant factor when considering modulus changes

as it is thought to be in the case of irradiation.

Figure 6.13 Change in Modulus of Brominated HOPG

Page 193: Using Intercalation to Simulate Irradiation Damage of

193

Table 6.4 Measured elastic constants of grade ZYH HOPG

Virgin HOPG

Literature Values

Brominated HOPG

Density 2.25E+03 Blakslee[171] Density 2.21E+03

Dimensions Strain Time of flight (s)

Modulus (Pa) Modulus (Pa) Dimensions Strain

Time of flight (s)

Modulus (Pa)

Zero time longitudinal 2.77E-05

C11 1.19E-02 0.00E+00 6.84E-07 6.83E+11 1.060E+12 1.20E-02 4.78E-03 2.34E-07 5.78E+12

C22 1.19E-02 0.00E+00 6.86E-07 6.80E+11 1.20E-02 4.86E-03 2.15E-07 6.91E+12

C33 3.80E-03 0.00E+00 1.08E-06 2.81E+10 3.65E+10 5.01E-03 3.19E-01 4.35E-07 6.10E+11

Zero time Shear

C44 3.80E-03 0.00E+00 1.06E-05 2.86E+08 0.18-0.35E+9 5.01E-03 3.19E-01 8.55E-06 3.34E+08

C66 1.19E-02 0.00E+00 1.12E-06 2.57E+11 1.20E-02 4.78E-03

Page 194: Using Intercalation to Simulate Irradiation Damage of

194

6.3.1.1. Error calculation of Modulus changes in brominated single crystal

graphite

The measurement of the elastic constants of brominated HOPG was a

difficult experiment. The elastic modulus values obtained are quite

high. The change in modulus for fully lithiated graphite has an

increase of 300% for C44, 150% for C33, and a 10% decrease for C11 as

calculated by Qi using DFT[114]. By comparison, this experiment

recorded changes in value nearly an order of magnitude larger.

Equation 6.10 shows the values to be measured to arrive at a value

for the young’s modulus all of which have associated errors.

2

321

t

l

lll

mE x

d

Equation 6. 10

where Ed is the dynamic Young’s modulus, m is the mass of the

sample lx is a length measurement and t is time. The fractional

certainties are given in Table 6.6. The uncertainties for all values

except time are very low. The requirement of PTFE rods however

greatly increases the error in the time measurement. Time is

measured across the sample and rods and by subtracting the time

measured across the PTFE rods with no sample present the time

through the sample is measured. This introduces a large error as the

pulse time across the rods is 2 orders of magnitude larger than

through the sample. Therefore a 5% error in the measurement of the

pulse arrival time can result in a 650% error in the calculated time for

the pulse to travel through the graphite sample.

Page 195: Using Intercalation to Simulate Irradiation Damage of

195

Table 6.5 Fractional Uncertainties in Young’s modulus measurement of HOPG

m 592.1

0005.0

l1 96.12

005.0

l2 98.11

005.0

l3 32.3

005.0

t

sample

samplePTFE

t

tt %5

By applying the uncertainties given in Table 6.6 to Equation 6.10 it

can be seen that the error in the time measurement is so much larger

than all other errors that they may be ignored. However as the

velocity measurement is squared the error is doubled. This means a

5% error in the time measurement propagates through to a 1300%

error in the value of Young’s modulus.

6.3.2. Modulus changes in brominated polycrystalline graphites

A series of experiments were carried out to measure the change of

sonic velocity in polycrystalline graphite. As Equation 6.2 shows, to

calculate the modulus the value of the sonic velocity is squared and it

is therefore the most dominant factor in the relationship. The change

in density of bromine with strain is introduced to the models using

the relevant equation, either Equation 6.5; Equation 6.6; or Equation

6.7. These experiments were performed on PGA and Gilsocarbon

graphite. In this section the data on change in modulus along with

Page 196: Using Intercalation to Simulate Irradiation Damage of

196

images of the samples are used to explain the trends. The results are

also compared with various graphite irradiation data.

6.3.2.1. PGA Against Grain

The results for the change in modulus of brominated PGA measured

perpendicular to extrusion (against grain) are shown in Figure 6.14.

There is significant scatter in the results though a general trend of

increasing modulus followed by a sudden decrease can be seen.

The initial results of 9.96GPa are high compared to the literature

value of 5.4GPa. It has not been possible to determine the reason for

this discrepancy. This work is interested in changes to microstructure

that cause a change in modulus, and with this data it is still possible

to relate trends in the data to microstructural changes. Later analysis

will present the data as the ratio E/E0. This ratio is commonly used

with irradiated nuclear graphites and this allows direct comparisons

between irradiated and intercalated modulus changes to be made.

The modulus of the graphite appears to have a bilinear trend. Initially

there is a modulus increase of approximately 19%. The modulus

increases until the intercalation has progressed to between 1% and

1.5% strain. The increase in modulus is most likely due to two

reasons; closure of Mrozowski cracks causing densification of the

crystallites, and an increase in the crystal modulus.

As the strain reaches between 1% and 1.6% there is a significant drop

in modulus. This coincides with the formation of large cracks across

Page 197: Using Intercalation to Simulate Irradiation Damage of

197

the sample as seen in Figure 6.15 and Figure 6.16. Referring back to

Figure 5.12, it is also the same region of strain when the increase in

the dimensional change rate ]/[

..

CBrd

CdD AG is observed and the drop in

density is measured, Figure 6.10.

It is postulated that the cracks are generated by well intercalated

filler particles generating large strains, the stresses of which must be

relieved through fractures propagating through the binder of the

microstructure. Investigations using confocal microscopy on the

debrominated specimen, Figure 6.11, show that the cracks run

through the binder and around the edges of the filler particle.

Coupled with knowledge gained from the synchrotron experiment in

Section 5.2 this tells us that the filler particles expand causing cracks

to propagate through the binder along the with grain direction. The

crack around the filler particle can be explained by the plastically

deformed binder matrix not returning to its original shape after the

filler particle has debrominated.

The rate of modulus drop off is lower for PGA Perp 3 than PGA Perp

1. This could be explained by differences in the cracks propagating

though the two samples. PGA Perp 1 which exhibits a sharp drop in

modulus, shows only one crack which has fractured right across the

middle of the specimen. In contrast, PGA Perp 3 shown in Figure 6.16

shows a number of smaller cracks running through the sample

contributing to a more gradual decrease in modulus.

Page 198: Using Intercalation to Simulate Irradiation Damage of

198

The area highlighted green in Figure 6.15 and Figure 6.16 show the

sample is a slightly darker colour in this region. This means that these

regions are at a slightly higher intercalation stage than the rest of the

sample[172], and therefore the measurement technique is

influencing the sample. Fortunately the damage appears to be to the

sides so it is only a small amount of laser damaged graphite that the

ultrasonic pulse must travel through.

Figure 6.14 Change in Young's modulus of brominated Pile Grade A cut perpendicular to extrusion (Against

Grain)

Page 199: Using Intercalation to Simulate Irradiation Damage of

199

Figure 6.15 Pile Grade A sample 1 cut against the grain (perpendicular to extrusion) after bromination

Figure 6.16 Pile Grade A sample 3 cut against the grain (perpendicular to extrusion) after bromination

6.3.2.2. PGA With Grain

The results of experiments carried out on Young’s modulus changes

in PGA cut parallel to extrusion (WG) are shown in Figure 6.17. The

initial results show good agreement with the literature data, 11.7GPa

[28] and 11.1GPa. The data shows a 25% increase in modulus at

around 1% to 1.5% strain. Again this is the region where the density

of graphite is shown to start increasing indicating porosity generation

in the binder matrix. The percentage increase in modulus is larger

with grain than against grain. This is presumably because the crystal

WG AG

WG AG

Page 200: Using Intercalation to Simulate Irradiation Damage of

200

modulus is able to reach a higher value before cracking in the

microstructure negates the crystal increase.

The data shows a similar trend to the PGA against grain samples,

though the drop in modulus is less dramatic. This can be explained by

the same process that is used to explain the difference in the drop in

modulus for the AG samples. Figure 6.18 shows that the cracks run

along the WG direction, this results in a more gradual increase in the

porosity in an orientation that will impede the sonic pulse.

Figure 6.17 Change in Young's Modulus of Brominated Pile Grade A cut Parallel to Extrusion

Page 201: Using Intercalation to Simulate Irradiation Damage of

201

Figure 6.18 Pile Grade A Sample after bromination

6.3.2.3. Gilsocarbon

The results for young’s modulus changes in brominated Gilsocarbon

are shown in Figure 6.19. The data shows an increase in modulus

followed by a subsequent drop when the intercalation strain reaches

1.5%. The initial increase is around 30%; this is the most significant

increase seen in all the graphites measured. The reason for this could

be the larger volume of accommodation porosity present in

Gilsocarbon as shown in Figure 5.25. This allows a larger percentage

increase in densification and therefore a bigger increase in modulus

before fracture of the binder becomes the dominant effect causing a

decrease in the modulus.

The microstructural cracks are shown in Figure 6.20, where the cracks

are much smaller in the Gilsocarbon samples than the PGA samples.

All that is left are small cracks surrounding the filler particle

presumed to have formed after debromination. This suggests that if

cracks are generated in Gilsocarbon they are much smaller. This

A

G

W

G

Page 202: Using Intercalation to Simulate Irradiation Damage of

202

would cause a more gradual reduction in the modulus as is seen in

the data.

Gilsocarbon sample 1 given in Figure 6.19 acts somewhat differently

to the other two samples. While two of the samples show a gradual

increase followed by a gradual decrease in modulus, sample 1 shows

a rapid increase in strain, followed by a sharp drop in modulus. This

may have been be caused by a problem with the laser displacement

detector. On the other hand, confocal microscopy images, Figure

6.20, show there is very little residual cracking in the debrominated

microstructure and the data point highlighted in red in Figure 6.19 is

the debrominated Young’s modulus value. This indicates that on

debromination the modulus returns to its original or a slightly higher

value than its virgin value. This suggests any cracking which opened

up in Gilsocarbon due to intercalation closes upon debromination

and therefore it is difficult to draw conclusions as to what happened

to this particular sample without having further data to verify the

laser displacement detector values

The modulus measurements are consistently lower than the

literature values with the virgin value of Gilsocarbon as 10.85GPa [28]

compared with measured value of 5.45GPa. As with the PGA against

grain sample; it was not possible to determine the cause of the

discrepancy between the two values.

Page 203: Using Intercalation to Simulate Irradiation Damage of

203

Change in modulus of Brominated Samples

1

2

3

4

5

6

7

8

9

-0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Strain (%)

Yo

un

g's

Mo

du

lus

(GP

a)

Gilso 1Gilso 4Gilso 2

Figure 6.19 Young's modulus changes in brominated Gilsocarbon

Figure 6.20 Laser confocal micrograph of debrominated Gilsocarbon showing reltivly little microstructural

cracking a) scale bar 100µm b) isometric image 1280 x 1280 µm

Rapid increase in strain

Page 204: Using Intercalation to Simulate Irradiation Damage of

204

6.3.3. Possible sources of Errors

There are many errors associated with this experiment as highlighted

by the scatter in the data.

There is an offset error to the unknowable zero time in the

experimental apparatus. This includes electronics delays and process

delays such as the time required for a laser pulse to generate the

ultrasonic pulse. Long samples are used to reduce this error.

The quality of the signal measured using the laser vibrometer and the

laser displacement detector is noisy even with filtering of the data.

This will introduce scatter to the data.

The quality of the vibrometer wave deteriorates as the surface of the

sample becomes more damaged by the intercalant. This reduces the

reflectivity of the sample making it harder to pick up the return

signal.

The modulus measurements carried out for the initial verification

experiment and the PGA parallel experiments all show quite good

agreement with the data. However, as discussed there are consistent

and repeatable errors between two of the data sets. This points to a

mix up of the samples. However, photos were taken of the samples

as soon as they were taken out of the experimental rig using both

standard photography and laser confocal microscopy and the

microstructural features match with what was expected of the

named sample.

Page 205: Using Intercalation to Simulate Irradiation Damage of

205

6.4. Comparison of modulus changes due to irradiation and bromination

Figure 6.21 to Figure 6.23 compare the effects of bromination with

irradiation on the change in Young’s modulus. It would have been

preferable to compare the bromination data with low temperature

irradiation data but unfortunately the data in the literature is too

sparse.

The irradiation data has been taken from Brocklehurst and Kelly’s

investigation into the Young’s modulus changes in polycrystalline

graphites[35]. The irradiation dose has been normalised to bromine

concentration by comparing the change in the c-axis of HOPG

brominated at room temperature and HOPG irradiated 430°C[36]

data to find the relationship between dose and bromination giving

1.1 x 1021EDN = 1 [Br/C].

The pinning term has been removed from the irradiation data. This is

because intercalation cannot simulate this aspect of irradiation

damage. This leaves the structural term which is thought to be due to

closure of porosity and later opening of new porosity[57].

It can be seen that all graphite grades follow the trend of increasing

then decreasing of modulus in both irradiation and bromination data

sets. As Chapter 5 shows, Mrozowski cracks are initially closed by

intercalation. As the intercalate concentration reaches higher levels

new porosity is opened as swelling filler particles fracture the

surrounding binder matrix. Density measurements show that the

Page 206: Using Intercalation to Simulate Irradiation Damage of

206

density of brominated samples drops at a similar strain to that where

a drop in modulus is measured, both occurrences can be explained by

an increase in porosity.

When normalizing the data sets to the crystal strain they receive in

the c axis, it can be seen that irradiated graphites can experience

more damage before measureable property changes are induced.

Irradiated samples also experience larger fractional changes in

modulus. This is due to differences between the irradiation and the

bromination damage mechanisms. Irradiation creep relaxes stresses

developed by expansion of filler particles. This allows a greater

change in the volume of the filler particles and greater densification

of filler particles[43] before fracture of the microstructure occurs.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 5E+21 1E+22 1.5E+22

[Br/C] adjusted for dose (n/cm-2

)

E/E

0

PGA perp 1

PGA perp 3

PGA perp irradiated

Figure 6.21 Comparison of changes in Young's modulus in PGA parallel to extrusion due to irradiation and

bromination

Page 207: Using Intercalation to Simulate Irradiation Damage of

207

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

0 5E+21 1E+22 1.5E+22 2E+22[Br/C] adjusted to dose (n/cm

-2)

(E/E

0)-

EP

PGA parallel 3

PGA parallel 5

PGA Irradiated at 600C

Figure 6.22 Comparison of changes in Young's modulus in PGA parallel to extrusion due to irradiation and

bromination

0.00E+00

5.00E-01

1.00E+00

1.50E+00

2.00E+00

2.50E+00

3.00E+00

3.50E+00

4.00E+00

0 5E+21 1E+22 1.5E+22 2E+22 2.5E+22

[Br/C] adjusted to Dose (n/cm-2

)

(E/E

0)-

Ep

Gilsocarbon 1

Gilsocarbon 2

Gilsocarbon 4

Gilsocarbon Irradiated at 430C

Gilsocarbon Irradiated at 600C

Figure 6.23 Comparison of changes in Young's modulus in Gilsocarbon due to irradiation and bromination

Page 208: Using Intercalation to Simulate Irradiation Damage of

208

6.5. Conclusions

This chapter shows that there are key similarities in the mechanisms

which induce changes in Young’s modulus in irradiated and

intercalated samples.

The chapter begins by describing the experimental technique

required to carry out Young’s modulus measurements under the

unique conditions imposed by bromination experiments. The chapter

describes the standard ASTM test technique using piezo transducers

and comparing it to a modified laser technique developed for this

work. Time is taken to describe how the experimental parameters

were defined for the laser technique. Results are presented which

compare the two techniques and it shown that the results are very

similar though there is a slight decrease in the velocity measurement

taken by laser and this is attributed to delays in the measurement

technique are unable to be accounted for.

The chapter goes on to show the results obtained for change in

modulus of single crystal and polycrystalline samples. The change in

modulus seen for single crystal graphite are very high and this is

attributed to large errors in the method used to measure the travel

time for an ultrasonic pulse.

The polycrystalline samples show an increase followed by a decrease

in the modulus of graphite as bromine intercalation increases. The

increase is attributed to a closure of porosity in the microstructure.

Page 209: Using Intercalation to Simulate Irradiation Damage of

209

The decrease is attributed to cracking induced in the microstructure

as heterogeneous strains become too large to be accommodated.

The polycrystalline results are compared to irradiation data and it can

be seen that the two methods both induce an increase followed by a

decrease in modulus and the reasons for the two changes are

thought to be the same namely an initial closure of porosity followed

by the onset of cracking in the microstructure. The magnitudes of

modulus change are much lower for brominated samples than

irradiated samples and this is attributed to fundamental differences

in the two damage processes namely the stress relief mechanism of

irradiation creep in the brominated samples.

Page 210: Using Intercalation to Simulate Irradiation Damage of

210

7. Conclusions and further work

7.1. Conclusions

An improved understanding of property changes is required to help

nuclear operators make safety cases for reactor life time extensions

and help designers create longer lasting more efficient reactors in the

future. The work in this thesis aims to help towards these goals by

providing a better understanding of the effect of internal strains on

material properties.

This work has used intercalation as a method to induce crystal strain.

Chapter 2 discusses intercalation of graphite in detail. The

fundamental point is that intercalation can induce an interplanar

strain of a similar magnitude to that caused by irradiation. This work

has utilised this effect to induce property changes in nuclear

graphite.

Bromine is a difficult substance to work with and so two unique

experimental rigs were designed to allow insitu property

measurements during intercalation. This is an advantage of using

intercalation over irradiation as insitu measurements would not be

possible with irradiation. The rigs allowed the first tomographic scans

to observe growth of graphite crystals due to interstitial damage. The

work also provided the first empirical measurements of modulus

changes in intercalated graphites.

Page 211: Using Intercalation to Simulate Irradiation Damage of

211

The results show that dimensional change goes through two stages.

The first stage affects the graphite crystallites at low [Br/C] ratios.

Intercalated bromine causes an expansion in the c-axis of crystals up

to a bromine concentration of 1.3 [Br/C]. The expansion causes a

closure of nanocracks; this is seen in both HOPG and polycrystalline

graphites. The closure of the cracks causes an increase in the density

of the crystallites and a corresponding increase in the density of bulk

graphite. The closure of cracks causes an increase in the modulus of

graphite polycrystalline graphites

At higher [Br/C] ratios, once crystal expansion has closed the

nanocracks the crystallites and the bulk structure start expanding at

an accelerated rate. Because intercalation initiates at the sample

surface a lot of intercalation initially occurs in the binder due to the

large surface area of binder exposed to a high partial pressure of

bromine and later on in the filler particles where it takes time for the

bromine to propogate through the crystal structure. The magnitudes

of strain induced in the microstructure by intercalation are shown by

tomography to close microcracks. As filler particle expansion occurs,

this generates strains which cause fracture in the binder matrix. This

causes fractures to propagate along the with grain direction as shown

by laser confocal microscopy and density measurements. The binder

fracturing corresponds to a decrease in the bulk modulus.

There are substantial differences between intercalation and

irradiation. There is probably no mechanism comparable to

Page 212: Using Intercalation to Simulate Irradiation Damage of

212

irradiation creep in graphite, meaning that any property changes

occur at lower levels of intercalation damage. There is no mechanism

to cause a-axis shrinkage which reduces the volume of irradiated

crystals. This prevents the calculation of an a-axis accommodation

factor as would be required to model irradiation damage suitably.

Despite the differences, it is still possible to answer some important

questions on irradiation properties at high dose. This work gives

evidence to support the theory that the initial increase in the

structural modulus term is due to closure of nanocracks and that the

later decrease in the term is due to the opening of cracks in the

binder region[57]. The tomographic investigations support the work

of Hall[51] by showing that large scale porosity will close given the

magnitudes of strain seen in a reactor. The work shows also that filler

particle expansion is a driving force behind property changes later in

core life including dimensional change and binder matrix fracture.

7.2. Future Work

It is recommended that future studies into the intercalation of

graphite as a method to simulate irradiation damage should focus on

the following areas.

The experiment to measure the change in modulus of HOPG was

unsatisfactory. It is recommended that a theoretical approach be

taken. DFT should be used to calculate the change in elastic constants

of intercalated crystals; this could build on the work of Yaya et al [71]

Page 213: Using Intercalation to Simulate Irradiation Damage of

213

who carried out DFT studies to find the interstitial positions of

bromine intercalated in graphite and graphene. The results of this

should be input into a finite element model to determine modulus

changes in crystallites.

It would be interesting to carry out high resolution powder diffraction

experiments. Such experiments could answer a number of questions.

How does intercalation differ between binder and filler materials? Is

there any reorientation of crystals? It could also be used to

accurately measure any a-axis contraction. If so, this would provide a

small insight into one of the mechanisms that may potentially occur

during irradiation creep[173].

If high resolution XRD were carried out and the value of Young’s

modulus of irradiated crystals were determined, it would be possible

to construct finite element meshes based on tomographic scans of

graphite, the results of which could be compared to matching

empirical data.

Given that some microstructural regions intercalate quicker than

others and therefore undergo modulus changes at different times,

the effect of this on the surrounding microstructure and bulk

material should be investigated.

Page 214: Using Intercalation to Simulate Irradiation Damage of

214

8. References

1. Telling, R.H. and M.I. Heggie, Radiation defects in graphite. Philosophical Magazine, 2007. 87(31): p. 4797 - 4846.

2. Argonne, N., Laboratory,. Atoms Forge a Scientific Revolution. [cited 2010]; Available from: http://www.anl.gov/Science_and_Technology/History/Anniversary_Frontiers/unisci.html.

3. Calder Hall Power Station, in The Engineer. 1956. 4. Association, W.N. Nuclear Power Reactors. Volume, 5. Abram, T. and S. Ion, Generation-IV nuclear power: A review of the state of the

science. Energy Policy, 2008. 36(12): p. 4323-4330. 6. F. Carré, P.Y., W.J. Lee, Y. Dong, Y. Tachibana and D. Petti. VHTR – ONGOING

INTERNATIONAL PROJECTS. in GIF Symposium. 2009. Paris, France. 7. Gen, I., Internationl, Forum,. 8. Coderre, J. Introduction to Ionizing Radiation. 2006 [cited 2012 6/6/2012]. 9. Krane, K.S., Introductory Nuclear Physics. 2 ed. 1987: John Wiley & Sons. 10. Marsden, B.J., Nuclear Graphite for High temperature Reactors. www.iaea.org,

2001. 11. Davidson, H.W., Losty, H.H.W., Mechanical Properties of Non-Metallic Brittle

Materials, ed. W.H. Walton. 1958: Butterworth. 12. Pierson, H.O., Handbook of Carbon, Graphite, Diamond and Fullerenes - Properties,

Processing and Applications. 1993: William Andrew Publishing/Noyes. 13. Walker, P.L., Chemistry and Physics Of Carbon,vol.2: A Series Of Advances. 1966:

Arnold. 14. Atamny, F., A. Baiker, and R. Schlögl, Atomic resolution of defects in graphite

studied by STM. Fresenius' Journal of Analytical Chemistry, 1997. 358(1): p. 344-348.

15. Chen, J., Introduction to Scanning Tunneling Microscopy. 2008: Oxford University Press.

16. Toshihiro Tsuji, H.I.a.K.Y., Observation of Anomalous Dislocation Behavior in Graphite Using Ultrasonic Atomic Force Microscopy. Japanese Journal of Applied Physics, 2002. 41: p. 832 - 835.

17. Haffenden, G., M. Heggie, and G. Sheehan. First Principles Calculations on the Thermal Properties of Irradiated Graphite in INGSM11. 2010. Eastbourne.

18. Jones, A.N., Hall, G. N., Joyce, M., Hodgkins, A., Wen, K., Marrow, T. J., Marsden, B. J., Microstructural characterisation of nuclear grade graphite. Journal of Nuclear Materials, 2008. 381(1-2): p. 152-157.

19. Mrozowski, S. Mechanical strength, thermal expansion and structure of cokes and carbons. in First Biennial Conference on Carbon. 1953. University of Buffalo, Buffalo.

20. Nightingale, R.E., Nuclear Graphite. 1962, New York: Academic Press. 21. W.V. Green, J.W.a.E.G.Z., High-temperature Creep of Polycrystalline Graphite.

Materials Science and Engineering A, 1970. V6: p. 199-211. 22. Wen, K.Y., T.J. Marrow, and B.J. Marsden, The microstructure of nuclear graphite

binders. Carbon, 2008. 46(1): p. 62-71. 23. Béghein, B., Mesnildot, Hiltmann, Melin. Nuclear Graphite NBG-17 and other

solutions for the nuclear industry. in INGSM11. 2010. Eastbourne. 24. Fischbach, D.B., Kinetics of Graphitization of a Petroleum Coke. Nature, 1963.

200(4913): p. 1281-1283.

Page 215: Using Intercalation to Simulate Irradiation Damage of

215

25. Kelly, B.T., Graphite--the most fascinating nuclear material. Carbon. Vol. 20. 1982. 3-11.

26. Thrower, P.A., Interstitial and vacancy loops in graphite irradiated at high temperatures. British Journal of Applied Physics, 1964. 15(10): p. 1153-1159.

27. Heerschap, M., P. Delavignette, and S. Amelinckx, Electron microscope study of interlamellar compounds of graphite with bromine, iodine monochloride and ferric chloride. Carbon, 1964. 1(3): p. 235-238, IN1-IN14, 239-243.

28. Department, f., Business, Enterprise, and, Regulatory, Reform, UK ENERGY IN BRIEF JULY 2008. 2008.

29. Ouseph, P.J., Observation of Prismatic Dislocations Loops in Graphite. Physica Status Solidi (a), 1998. 169(1): p. 25-32.

30. Thrower, P.A., The Study of Defects in Graphite by Transmission Electron Microscopy. Chemistry and Physics of Carbon, ed. P.L. Walker. Vol. 5. 1969.

31. Henson, R.W., A.J. Perks, and J.H.W. Simmons, Lattice parameter and dimensional changes in graphite irradiated between 300 and 1350[deg]C. Carbon, 1968. 6(6): p. 789-804.

32. Heggie, M.I., Suarez-Martinez, I., Davidson, C., Haffenden, G., Buckle, ruck and tuck: A proposed new model for the response of graphite to neutron irradiation. Journal of Nuclear Materials. In Press, Corrected Proof.

33. Kelly, B.T., W.H. Martin, and P.T. Nettley, Dimensional Changes in Pyrolytic Graphite under Fast-Neutron Irradiation. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1966. 260(1109): p. 37-49.

34. Prince, N., C.R. Simons, and J.D. Hart. Moderator Structure Design In the UK - A historical Survey. in Graphite Structures for Nuclear reactors. 1972. London.

35. Brocklehurst, J.E. and B.T. Kelly, Analysis of the dimensional changes and structural changes in polycrystalline graphite under fast neutron irradiation. Carbon, 1993. 31(1): p. 155-178.

36. Brocklehurst, J.E. and B.T. Kelly, The dimensional changes of highly-oriented pyrolytic graphite irradiated with fast neutrons at 430°C and 600°C. Carbon, 1993. 31(1): p. 179-183.

37. Kinchin., G.H. The effects of irradiation on graphite. in Proceedings of the International Conference on the Peaceful Uses of Atomic Energy,. 1956. 38. Bridge, H., B.T. Kelly, and P.T. Nettley, Effect of high-flux fast-neutron irradiation on

the physical properties of graphite. Carbon, 1964. 2(1): p. 83-93. 39. Birch, M. and J.E. Brocklehurst, A review of the behaviour of graphite under the

conditions appropriate for protection of the first wall of a fusion reactor. 1987, [Risley] Warrington: Risley Nuclear Power Development Establishment.

40. Morgan, W.C., Notes on the two-phase model for radiation-induced dimensional changes in graphite. Carbon, 1966. 4(2): p. 215-222.

41. Neighbour, G.B., Modelling of dimensional changes in irradiated nuclear graphites. Journal of Physics D: Applied Physics, 2000.

42. Tsang, D.K.L. and B.J. Marsden, Effects of dimensional change strain in nuclear graphite component stress analysis. Nuclear Engineering and Design, 2007. 237(9): p. 897-904.

43. Burchell, T.D., Irradiation induced creep behavior of H-451 graphite. Journal of Nuclear Materials, 2008. 381(1-2): p. 46-54.

44. Simmons, J.H.W. The effects of irradiation on the mechanical properties of graphite. in Third Conference on Carbon. 1959. Buffalo: Pergamon Press, New York.

45. Tsang, D.K.L., et al., Graphite thermal expansion relationship for different temperature ranges. Carbon, 2005. 43(14): p. 2902-2906.

Page 216: Using Intercalation to Simulate Irradiation Damage of

216

46. Sutton, A.L. and V.C. Howard, The role of porosity in the accommodation of thermal expansion in graphite. Journal of Nuclear Materials, 1962. 7(1): p. 58.

47. Brocklehurst, J.E. and J.C. Weeks, Dimensional changes in graphite: The relationship between those produced by absorption of bromine and those produced by irradiation. Journal of Nuclear Materials, 1963. 9(2): p. 197-210.

48. Martin, W.H. and J.E. Brocklehurst, The thermal expansion behaviour of pyrolytic graphite-bromine residue compounds. Carbon, 1964. 1(2): p. 133-134, IN3, 135-141.

49. Hall, G., Marsden, B. J., Fok, S. L., Smart, J., The relationship between irradiation induced dimensional change and the coefficient of thermal expansion: a modified Simmons relationship. Nuclear Engineering and Design, 2003. 222(2-3): p. 319.

50. Hall, G., B.J. Marsden, and S.L. Fok, The microstructural modelling of nuclear grade graphite. Journal of Nuclear Materials, 2006. 353(1-2): p. 12-18.

51. Hall, G.N., MICROSTRUCTURAL MODELLING OF NUCLEAR GRADE GRAPHITE, in Mechanical Engineering. 2004, University of Manchester.

52. ASTM E111 - 04(2010) Standard Test Method for Young's Modulus, Tangent Modulus, and Chord Modulus, ed. A. International. 2010.

53. Oku, T. and M. Eto, Relation between static and dynamic Young's modulus of nuclear graphites and carbon. Nuclear Engineering and Design, 1993. 143(2-3): p. 239-243.

54. Yoda, S. and M. Eto, The tensile deformation behavior of nuclear-grade isotropic graphite posterior to hydrostatic loading. Journal of Nuclear Materials, 1983. 118(2–3): p. 214-219.

55. Barron, T.H.K., Klein, M. L., Second-order elastic constants of a solid under stress. Proceedings of the Physical Society, 1965. 85(3): p. 523.

56. Seldin, E.J. and C.W. Nezbeda, Elastic Constants and Electron-Microscope Observations of Neutron-Irradiated Compression-Annealed Pyrolytic and Single-Crystal Graphite. Journal of Applied Physics, 1970. 41(8): p. 3389-3400.

57. Kelly, B.T., The elastic constants of polycrystalline carbons and graphites. Philosophical Magazine, 1964. 9(101): p. 721 - 737.

58. M. Metcalf, J.F.B.P. The Effects of Thermal Annealing on the Mechanical Properties of PGA Graphite. in Managment of Ageing in Graphite Reactor Cores. 2005. Cardiff.

59. Oku, T., K. Fujisaki, and M. Eto, Irradiation creep properties of a near-isotropic graphite. Journal of Nuclear Materials, 1988. 152(2-3): p. 225-234.

60. Kelly, B.T., The physics of graphite. 1981, London: Applied Science Publishers. 61. Kelly, B.T., Irradiation Damage to Solids. 1966: Pergamon Press. 232. 62. Telling, R.H., Ewels, C. P., El-Barbary, A. A., Heggie, M. I., Wigner defects bridge the

graphite gap. Nature Materials, 2003. 2(5): p. 333-337. 63. Brocklehurst, J.E., Dimensional Changes in Graphite: The Relationship Between

Those Produced by Absorption of Bromine and Those Produced by Thermal Expansion and by Fast Neutron Irradiation. 1966, University of Manchester: Manchester. p. 49.

64. A. Weiss, A Secret of Chinese Porcelain Manufacutre. Angewandte Chemie International Edition in English, 1963. 2: p. 697-703.

65. Vandiver, P.B., Soffer, O., Klima, B., Svoboda, J., The Origins of Ceramic Technology at Dolni Věstonice, Czechoslovakia. Science, 1989. 246(4933): p. 1002-1008.

66. M. S. Whittingham, A.J.J., Intercalation Chemistry. Materials Science Series. 1982: Academic Press. 587.

67. Hofmann, U., Frenzel, A., Über permutoide Reaktionen des Graphits. Zeitschrift für Elektrochemie und angewandte physikalische Chemie, 1931. 37(8-9): p. 613-618.

68. A. Schleede, M.W., Z. phys. Chem.,, 1932. 1(B 1): p. 18.

Page 217: Using Intercalation to Simulate Irradiation Damage of

217

69. Bedioui, F., Zeolite-encapsulated and clay-intercalated metal porphyrin, phthalocyanine and Schiff-base complexes as models for biomimetic oxidation catalysts: an overview. Coordination Chemistry Reviews, 1995. 144: p. 39-68.

70. Dresselhaus, M.S., Dresselhaus, G., Intercalation compounds of graphite. Advances in Physics, 2002. 51(1): p. 1 - 186.

71. Yaya A., E.C.P., Suarez-Martinez I., Wagner Ph. Lefrant S., Okotrub A., Bulusheva L., Briddon P. R., Bromination of graphene and graphite. Physical Review B. 83(4): p. 045411.

72. Kozlov, M.E., Spin-glass behaviour of nanocrystalline diamond intercalated with potassium. Journal of Physics: Condensed Matter, 1997. 9(39): p. 8325.

73. Rosseinsky, M.J., Fullerene intercalation chemistry. Journal of Materials Chemistry, 1995. 5(10): p. 1497-1513.

74. Zhou, O., Gao, B., Bower, C., Fleming, L., Shimoda, H., Structure and Electrochemical Properties of Carbon Nanotube Intercalation Compounds. Molecular Crystals and Liquid Crystals Science and Technology. Section A. Molecular Crystals and Liquid Crystals, 2000. 340: p. 541 - 546.

75. Weller, T.E., Ellerby, M., Saxena, S.S., Smith, R.P., Skipper, N.T., Superconductivity in the intercalated graphite compounds C6Yb and C6Ca. Nat Phys, 2005. 1(1): p. 39-41.

76. T. Enoki, M.S., M. Endo, Graphite Intercalation Compounds and Applications. 2003: Oxford University Press.

77. Srinivas, G., Lovell, A., Skipper, N. T., Bennington, S. M., Kurban, Z., Smith, R. I., Ammonia absorption in calcium graphite intercalation compound: in situ neutron diffraction, Raman spectroscopy and magnetization. Physical Chemistry Chemical Physics. 12(23): p. 6253-6259.

78. Beguin, F., Setton, R., Facchini, L., Legrand, A.P., Merle, G., Mai, C., Structure and properties of KC24(Bz)2, A graphite-potassium-benzene intercalation compound. Synthetic Metals, 1980. 2(3-4): p. 161-170.

79. Rudorff, W., The solution of bromine in the crystal lattice of graphite: bromine graphite. Z Anorg. Allg. Chem, 1941.

80. Herold, A., Réflexions sur la synthèse des composés lamellaires. Materials Science and Engineering, 1977. 31: p. 1-16.

81. Sasa, T., Takahashi, Y., Mukaibo, T., Crystal structure of graphite bromine lamellar compounds. Carbon, 1971. 9(4): p. 407-416.

82. Marie, J., Mering, J., Fixation of bromine on carbons of varying degrees of graphetisation. Proc. 3rd Biennial Carbon Conf, 1957: p. 337.

83. Eeles, W.T. and J.A. Turnbull, The Crystal Structure of Graphite-Bromine Compounds. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1965. 283(1393): p. 179-193.

84. Miyazaki, T., Oshida, K., Endo, M., Dresselhaus, M.S., Application of Image Analysis for TEM Image of Acceptor Graphite Intercalation Compound. Molecular Crystals and Liquid Crystals Science and Technology. Section A. Molecular Crystals and Liquid Crystals, 2000. 340: p. 241 - 246.

85. Nixon, D.E., G.S. Parry, and A.R. Ubbelohde, Order-Disorder Transformations in Graphite Nitrates. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1966. 291(1426): p. 324-339.

86. Ghosh, D. and D.D.L. Chung, Two-dimensional structure of bromine intercalated graphite. Materials Research Bulletin, 1983. 18(10): p. 1179-1187.

87. Mochrie, S.G.J., et al., Novel melting transition in a two-dimensional stripe-domain system. Physical Review Letters, 1987. 58(7): p. 690.

Page 218: Using Intercalation to Simulate Irradiation Damage of

218

88. Wiesler, D.G., H. Zabel, and S.M. Shapiro, Two dimensional XY type magnetism in intercalated graphite: an elastic and inelastic neutron scattering study. Zeitschrift für Physik B Condensed Matter, 1994. 93(3): p. 277-297.

89. Simon, C., Rosenman, I., Batallan, F., Pepy, G., Lauter, H., Neutron diffraction on bromine intercalated in graphite. Synthetic Metals, 1988. 23(1-4): p. 147-153.

90. Eklund, P.C., Kambe, N., Dresselhaus, G., Dresselhaus, M. S., In-plane intercalate lattice modes in graphite-bromine using Raman spectroscopy. Physical Review B, 1978. 18(12): p. 7069.

91. Simon, C., Batallan, F., Rosenman, I., Pepy, G., Lauter, H., The anisotropy of the C44 shear constant in bromine-graphite intercalation compound: A precursor effect of the anisotropic melting. Physica B+C. 136(1-3): p. 15-17.

92. Reynolds, W.N. and P.A. Thrower, The nucleation of radiation damage in graphite. Philosophical Magazine, 1965. 12(117): p. 573 - 593.

93. Hennig, G., The Properties of the Interstitial Compounds of Graphite. III. The Electrical Properties of the Halogen Compounds of Graphite. The Journal of Chemical Physics, 1952. 20(9): p. 1443-1447.

94. Hooley, J.G. and J.L. Smee, The mechanism of the bromination of graphite. Carbon, 1964. 2(2): p. 135-138.

95. Takai K., K.H., Sato H., Enoki T., Bromine-adsorption-induced change in the electronic and magnetic properties of nanographite network systems. Physical Review B, 2006. 73(3): p. 035435.

96. Hooley, J.G., Physical chemistry and mechanism of intercalation in graphite. Materials Science and Engineering, 1977. 31: p. 17-24.

97. Safran, S.A. and D.R. Hamann, Coherency strains and staging in graphite intercalation compounds. Physica B+C, 1980. 99(1-4): p. 469-472.

98. Kirczenow, G., Interference Phenomena in the Theory of Daumas-Hérold Domain Walls. Physical Review Letters, 1982. 49(25): p. 1853.

99. Axdal S. H., C.D.D.L., A theory for the kinetics of intercalation of graphite. Carbon, 1987. 25(3): p. 377-389.

100. Clarke, R., N. Wada, and S.A. Solin, Pressure-Induced Staging Transition in KC24. Physical Review Letters, 1980. 44(24): p. 1616.

101. Thomas J. M., M.G.R., Schlögl R. F., Boehm H. P., Direct imaging of a graphite intercalate: Evidence of interpenetration of [`]stages' in graphite: Ferric chloride. Materials Research Bulletin, 1980. 15(5): p. 671-676.

102. Evans, E.L. and J.M. Thomas, Ultramicrostructural characteristics of some intercalates of graphite: An electron microscopic study. Journal of Solid State Chemistry, 1975. 14(1): p. 99-111.

103. Timp, G., Dresselhaus, M. S. , The ultramicrostructure of commensurate graphite intercalation compounds. Journal of Physics C: Solid State Physics, 1984. 17(15): p. 2641.

104. Levi-Setti R., C.G., Wang Y. L., Parker N. W., Mittleman R., Hwang D. M., High-Resolution Scanning-Ion-Microprobe Study of Graphite and its Intercalation Compounds. Physical Review Letters, 1985. 54(24): p. 2615.

105. Axdal, S.H., Chung, D. D. L., Kinetics and thermodynamics of intercalation of bromine in graphite--I. Experimental. Carbon, 1987. 25(2): p. 191-210.

106. Blackman, L.C., G. Saunders, and A.R. Ubbelohde, Defect Structure and Properties of Pyrolytic Carbons. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 1961. 264(1316): p. 19-40.

107. Hooley, J.G., W.P. Garby, and J. Valentin, The effect of sample shape on the bromination of graphite. Carbon, 1965. 3(1): p. 7-16.

Page 219: Using Intercalation to Simulate Irradiation Damage of

219

108. Khvostikova O., H.H., Wendrock H., Gemming T., Thomas J., Ehrenberg H., Effect of the microstructure on the intercalation and exfoliation behaviour of graphite. Journal of Materials Science. 46(8): p. 2422-2430.

109. Hooley, J.G., Intercalation Isotherms on Graphites, in Chemistry and Physics of Carbon, P.L. Walker, Editor. 1969, Marcel Dekker, Inc: New York. p. 321-372.

110. Saunders, G.A., A.R. Ubbelohde, and D.A. Young, The Formation of Graphite/Bromine. I. Hysteresis of Bromine Insertion Between the Carbon Hexagon Layers. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1963. 271(1347): p. 499-511.

111. Simmons, J.H.W., A relation between thermal expansion and dimensional change for polycrystalline graphite, in UKAEA Report. 1961, Atomic Energy Research Establishment.

112. Shearing, P.R., Howard, L. E., Jørgensen, P. S., Brandon, N. P., Harris, S. J., Characterization of the 3-dimensional microstructure of a graphite negative electrode from a Li-ion battery. Electrochemistry Communications. 12(3): p. 374-377.

113. Qi, Y. and S.J. Harris, In Situ Observation of Strains during Lithiation of a Graphite Electrode. Journal of The Electrochemical Society. 157(6): p. A741-A747.

114. Qi Y., G.H., Hector L. G., Timmons A., Threefold Increase in the Young's Modulus of Graphite Negative Electrode during Lithium Intercalation. Journal of The Electrochemical Society. 157(5): p. A558-A566.

115. Taylor, R., Brown, R. G., Gilchrist, K., Hall, E., Hodds, A. T., Kelly, B. T., Morris F., The mechanical properties of reactor graphite. Carbon, 1967. 5(5): p. 519-531.

116. Gaier, J.R., Lake, M. L., Moinuddin, A., Marabito, M., Properties of hybrid CVD/PAN graphite fibers and their bromine intercalation compounds. Carbon, 1992. 30(3): p. 345-349.

117. Jaworske, D.A., R.D. Vannucci, and R. Zinolabedini, Mechanical and Electrical Properties of Graphite Fiber-Epoxy Composites Made from Pristine and Bromine Intercalated Fibers. Journal of Composite Materials, 1987. 21(6): p. 580-592.

118. McBain, J.W., Bakr, A. M., A NEW SORPTION BALANCE. Journal of the American Chemical Society, 1926. 48(3): p. 690-695.

119. Czichos H., S.T., Smith L., Springer Handbook of Materials Measurement Methods. Vol. 1. 2006: Springer.

120. Buzug, T., Computed Tomography From Photon Statistics to Modern Cone-Beam CT. 2008: Springer.

121. Attwood, D.T. Introduction to Synchrotron Radiation. 2007 [cited 19/08/2011]; Lecture Notes]. Available from: http://ast.coe.berkeley.edu/srms/.

122. Herman, G.T., Fundamentals of Computerized Tomography Image Reconstruction from Projections. 2009: Springer.

123. Ioka, I. and S. Yoda, Measurements of the density profile in oxidized graphite by X-ray computed tomography. Journal of Nuclear Materials, 1988. 151(2): p. 202-208.

124. Babout, L., Marsden, B. J., Mummery, P. M., Marrow, T. J., Three-dimensional characterization and thermal property modelling of thermally oxidized nuclear graphite. Acta Materialia, 2008. 56(16): p. 4242-4254.

125. D. James, G.H., P. Mummery, J. Marrow. Application of Computerised Tomography and Image Correlation to the Microstructural Modelling of Gilsocarbon. in INGSM 10. 2009. West Yellowstone Idaho.

126. M. Stampanoni, A.G., A. Isenegger, G. Mikuljan, Q. Chen, A. Bertrand, and R.B. S. Henein, U. Frommherz, P. B¨ohler, D. Meister, M. Lange and R. Abela. Trends in synchrotron-based tomographic imaging: the SLS experience. in Developments in X-Ray Tomography V. 2006.

Page 220: Using Intercalation to Simulate Irradiation Damage of

220

127. Stutton. M, A., Orteu. J, J, Schreier. H, W,, Image Correlation for Shape, Motion and Deformation Measurements. 2009: Springer.

128. Schreier, H., Orteu, J.J., Sutton, M.A., Image Correlation for Shape, Motion and Deformation Measurements Basic Concepts,Theory and Applications. Vol. 1. 2009: Springer.

129. Becker, T.H., T.J. Marrow, and R.B. Tait, Damage, crack growth and fracture characteristics of nuclear grade graphite using the Double Torsion technique. Journal of Nuclear Materials, 2011. 414(1): p. 32-43.

130. Joyce, M.R., Marrow, T. J., Mummery, P., Marsden, B. J., Observation of microstructure deformation and damage in nuclear graphite. Engineering Fracture Mechanics, 2008. 75(12): p. 3633-3645.

131. Li, H., J. Duff, and T.J. Marrow, In-Situ Observation of Crack Nucleation in Nuclear Graphite by Digital Image Correlation. ASME Conference Proceedings, 2008. 2008(48296): p. 813-820.

132. Timoshenko, S.P., Theory of Elasticity. 1970: McGraw & Hill. 133. ASTM C769 - 98(2005) "Standard Test Method for Sonic Velocity in Manufactured

Carbon and Graphite Materials for Use in Obtaining Young's Modulus". www.astm.org, ed. A. International. 2005.

134. ASTM C769 - 09 "Standard Test Method for Sonic Velocity in Manufactured Carbon and Graphite Materials for Use in Obtaining Young's Modulus". www.astm.org, ed. A. International. 2009.

135. Dutton, B., Dewhurst, R. J., Graphite anisotropy measurements using laser-generated ultrasound. Journal of Optics A: Pure and Applied Optics, 2007. 9(6): p. S7.

136. Scruby, C.B., Drain, L.E., Laser Ultrasonics: Techniques and applications. 1990: Adam Hilger.

137. K. Gatzwiller, K.B.G. Practical Aspects of Successful Laser Doppler Vibrometry Based Measurements. in 2003 IMAC-XXI: Conference & Exposition on Structural Dynamics. 2003. Florida.

138. Vass, J., Smid, R., Randall, R. B., Sovka, P., Cristalli, C., Torcianti, B., Avoidance of speckle noise in laser vibrometry by the use of kurtosis ratio: Application to mechanical fault diagnostics. Mechanical Systems and Signal Processing, 2008. 22(3): p. 647-671.

139. Smallman, R.E. and R.J. Bishop, Modern Physical Metallurgy and Materials Engineering - Science, Process, Applications (6th Edition). 1999, Elsevier.

140. Bragg, W., The intensity of reflexion of X-rays by crystals. Acta Crystallographica Section A, 1969. 25(1): p. 3-11.

141. Cullity, B.D., Elements of X-ray Diffraction. 1956: Addison-Wesley. 142. Scherrer. P., Bestimmung der Grösse und der inneren Struktur von Kolloidteilchen

mittels Röntgenstrahlen. Nachr. Ges. Wiss. Göttingen, 1918. 26. 143. O. Engler., V.R., Introduction to Texture Analysis - Macrotexture, Microtexture and

Orientation Mapping. 2 ed. 2010: CRC Press. 144. Bacon, G.E., A method for determining the degree of orientation of graphite.

Journal of Applied Chemistry, 1956. 6(11): p. 477-481. 145. Micrometrics, Micrometics Analysis System Getting Started. 2006. 146. Perrin, S., Study of Dimensional changes in various types of nuclear graphites by

Bromination and the relationship with the coefficient of thermal expansion, in School of Materials. 2007, University of Manchester: Manchester.

147. Sasa, T., Vapor pressure and hysteresis in the graphite-bromine system. Carbon, 1973. 11(5): p. 497-503.

148. PSI. [cited 14/08/11]; Available from: http://www.psi.ch/sls/tomcat/source.

Page 221: Using Intercalation to Simulate Irradiation Damage of

221

149. NIST. [cited 14/08/2001]; Available from: http://physics.nist.gov/PhysRefData/XrayMassCoef/tab4.html.

150. Russo, R.E., Mao, X. L., Borisov, O. V., Liu, Haichen, Influence of wavelength on fractionation in laser ablation ICP-MS. Journal of Analytical Atomic Spectrometry, 2000. 15(9): p. 1115-1120.

151. EdmundOptics. [cited 14/08/11]; Available from: http://www.edmundoptics.com. 152. Standring, J. and B.W. Ashton, The effect of radiolytic oxidation by carbon dioxide

on the porosity of graphite. Carbon, 1965. 3(2): p. 157-165. 153. Commons, W. FWHM.svg. 2012 [cited; Available from:

http://en.wikipedia.org/wiki/File:FWHM.svg. 154. Thorne, R.P., Cowen, H.C., Kelly, B.T., Improved Graphites. The Reactor: The Journal

of the Reactor Group, UKAEA, 1964. 3(3): p. 5-8. 155. T. Beirne, R.H.K., I.B. Mason, Some Physical Properties of Commercial Graphites.

UKAEA Research Group Memorandum, AERE-M8111, 1961. 156. R. Takahashi, M.T., S. Maruki, H. Ueda, T. Yamamoto. Investigation of morphology

and impurity of nuclear-grade graphite, and leaching mechanism of carbon-14. in IAEA Technical Committee Meeting on "Nuclear Graphite Waste Management". 1999. Manchester, United Kingdom: IAEA.

157. Kelly, B.T., The irradiation of graphite in materials testing reactors. The Reactor: The Journal of the Reactor Group, UKAEA, 1966. 5(3): p. 7-9.

158. Engle, G.B., Density and structural distributions in artificial graphites. Carbon, 1970. 8(4): p. 485-490, IN15-IN16, 491-495.

159. Aldrich, S., Analytical Reagant Grade Bromine Saftey Data Sheet. 2011, Sigma Aldrich.

160. Ceramics, M. [cited 14/08/11]; Available from: www.momentive.com. 161. Eklund, P.C., Kambe, N., Dresselhaus, G., Dresselhaus, M. S., In-plane intercalate

lattice modes in graphite-bromine using Raman spectroscopy. Physical Review B, 1978. 18(12): p. 7069-7079.

162. Kelly, B.T., GRAPHITE - THE MOST FASCINATING NUCLEAR MATERIAL. Carbon, 1982. 20(1): p. 3-11.

163. Sutton, A.L., Howard, V. C., Rate of porosity in the accommodation of thermal expansion in graphite. Carbon, 1964. 1(3): p. 367-368.

164. M. Stampanoni, A.G., A. Isenegger, G. Mikuljan, Q. Chen, A. Bertrand, S. Henein, R. Betemps, U. Frommherz, P. Bohler, D. Meister, M. Lange and R. Abela. Trends in synchrotron-based tomographic imaging: the SLS experience. in Proc. of SPIE. 2006.

165. Louis, L., Wong, T.F., Baud, P., Imaging strain localization by X-ray radiography and digital image correlation: Deformation bands in Rothbach sandstone. Journal of Structural Geology, 2007. 29(1): p. 129-140.

166. Kumar, U., Ramakrishna, G. S., Pendharkar, A. S., Kailas, S., A statistical correction method for minimization of systemic artefact in a continuous-rotate X-ray based industrial CT system. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2003. 515(3): p. 829-839.

167. W. Bodel, A.N.J., T. J. Marrow, B. J. Marsden, M. Schmidt, P. Mummery, L. Poulter, M. Lamb, C. Atkin. Dynamic Young's Modulus Measurements of UK Nuclear Graphite. in Modelling and Measuring Reactor Core Graphite Properties and Performance. 2011. Birmingham.

168. J. N. Caron, Y.Y., J. B. Mehl and K. V. Steiner, THermoelastic and Ablative Laser Generation of Ultrasonic Waveforms in Graphite/Polymer Composite Materials. Journal of Applied Physics, 1998.

Page 222: Using Intercalation to Simulate Irradiation Damage of

222

169. Corbett, R.E. and R.J. Dewhurst, PROPAGATION OF LASER-GENERATED ACOUSTIC PULSES IN AGR GRAPHITE. Nondestructive Testing and Evaluation, 1990. 5(2-3): p. 121-134.

170. Preston, S.D. and B.J. Marsden, Changes in the coefficient of thermal expansion in stressed Gilsocarbon graphite. Carbon, 2006. 44(7): p. 1250-1257.

171. Blakslee, O.L., Proctor, D. G., Seldin, E. J., Spence, G. B., Weng, T., Elastic Constants of Compression‐Annealed Pyrolytic Graphite. Journal of Applied Physics, 1970. 41(8): p. 3373-3382.

172. M. S. Dresselhaus, G.D., Intercalated Layered Materials, ed. F. Levy. 1979: Springer. 173. Burchell, T., K. Murty, and J. Eapen, Irradiation induced creep of graphite. JOM

Journal of the Minerals, Metals and Materials Society. 62(9): p. 93-99.