Академический Документы
Профессиональный Документы
Культура Документы
MATERIALS SCIENCE
M17: Nuclear Materials (Lectures 18)
Name............................. College..........................
Dr J. H. Gwynne
Lent Term 2013-14
III
Lent 2014
INTRODUCTION
In existing technologies for nuclear power generation there are many performance, reliability
and safety issues centred on the materials in use. In particular, materials degradation is the
principal obstacle to extending the lifetime of an existing plant. In almost every case, materials
issues pose the greatest challenge in bringing next-generation reactor designs to fruition. Of
course, many of the challenges in nuclear power generation, however severe, are of a general
kind; examples are high temperatures and corrosive environments.
The focus of this part of the course will be the specific effects of radiation on structural
materials and in the reactor core. The effects of radiation can be dramatic: change in shape,
swelling by some tens of percent, hardening (more than five-fold), drastic embrittlement and
reduction in ductility, and accelerated corrosion effects such as environmentally induced
cracking.
Useful books
K.L. Murty & I. Charit, An Introduction to Nuclear Materials Wiley (2013)
B.R.T. Frost (ed.), Nuclear Materials (Vols 10A & 10B, Materials Science & Technology),
VCH (1994)
B.M. Ma, Nuclear Reactor Materials and Applications, Van Nostrand (1983)
G.S. Was, Fundamentals of Radiation Materials Science, Springer (2007)
Additional Resources
-1-
Lent 2014
Contents
1. Introduction
1.1 Fundamentals
1.1.1 Notation
1.1.2 Binding energy
1.1.3 Types of radiation and radioactive decay
1.1.4 Exponential decay and half-life
1.1.5 Neutron classification
5
5
5
5
6
7
8
9
10
10
10
10
12
14
15
15
16
18
19
20
20
21
21
22
22
2. Radiation Damage
2.1 Introduction
2.2 Knock-on atoms and displacement cascades
2.3 Dislocation loops
2.4 Nucleation of cavities and voids
23
23
23
25
27
3. Nuclear fuels
3.1 Introduction
3.2 Metallic fuels
3.2.1 Uranium
3.2.1.1 Structure
3.2.1.2 Thermal expansion and thermal cycling growth
3.2.1.3 Thermal conductivity
28
28
30
30
30
31
32
-2-
Lent 2014
33
34
34
35
37
38
39
40
41
42
43
46
4. Cladding
4.1 Introduction
4.2 Austenitic stainless steels
4.2.1 Helium production and void swelling
4.2.1.1 Minimising void swelling
4.2.2 Inverse Kirkendall effect
4.2.3 Dislocation densities
4.2.4 Creep
4.2.5 Mechanical properties
4.2.6 Summary
4.3 Ferritic alloys
4.4 Zirconium alloys
4.4.1 Structure
4.4.2 Effects of irradiation
4.4.2.1 Irradiation growth
4.4.2.2 Irradiation creep
4.4.2.3 Mechanical properties
4.4.2.4 Corrosion
47
47
47
48
52
53
54
55
56
58
59
60
61
61
62
63
63
64
5. Moderators
5.1 Introduction
5.2 Graphite
5.2.1 Structure
5.2.2 Effect of irradiation on properties
5.2.3 Wigner energy
5.3 Other solid moderators
5.3.1 ZrH
5.3.2 Beryllium
5.4 Liquid moderators
65
65
67
67
68
69
70
70
70
70
-3-
Lent 2014
6. Control rods
6.1 Introduction
6.2 Materials used for control rods
71
71
72
73
73
73
77
78
78
78
78
80
81
Glossary
Abbreviations
82
84
Question sheet 1
Question sheet 2
Examples class
85
87
88
-4-
Lent 2014
1. Introduction
1.1
Fundamentals
1.1.1 Notation
A nuclide (any nucleus, any isotope of any element) can be represented as:
where:
A is the mass number (number of nucleons, i.e. number of protons and neutrons)
Z is the atomic number (number of protons in the nuclide)
N is the neutron number
A=Z+N
-5-
Lent 2014
or .
The most common source of alpha particles is the alpha decay of heavy atoms: when an atom
emits an alpha particle, its mass number decreases by four and its atomic number decreases by
two. This is an example of transmutation: the conversion of one chemical element or isotope
into another. For example, when
undergoes alpha decay, it emits an alpha particle and
forms
:
Beta radiation
Beta particles are high-energy, high-speed electrons ( ) or positrons ( ) emitted by certain
radioactive nuclei. The production of beta particles is known as beta decay, which also
involves transmutation.
If an unstable nucleus has an excess of neutrons, it may undergo
decay, in which a neutron
is converted into a proton and an electron (and an antineutrino). An example is the decay of
into
:
into
Gamma radiation
Gamma radiation, , is high-frequency (and therefore high-energy) electromagnetic radiation.
Gamma rays are produced during gamma decay, which occurs following alpha or beta decay.
The daughter nucleus is usually left in an excited state, so it can move to a lower energy state
by emitting a gamma ray.
-6-
Lent 2014
chance that an atom will have undergone nuclear decay. It can also be thought of as the time
after which half the radioactive atoms will have decayed. It therefore follows that:
Therefore,
Note that N(t) can also be written as:
The half-life varies depending on the isotope, and is usually determined experimentally. Halflives can vary from 10-24 s to 10+30 s (54 orders of magnitude!). For a list of half-lives, see:
http://en.wikipedia.org/wiki/List_of_isotopes_by_half-life
-7-
Lent 2014
Energy
Cold neutrons
<0.003 eV
0.003 0.4 eV
0.4 100 eV
Fast neutrons
>10 MeV
Generally, thermal neutrons are associated with a kinetic energy of ~0.025 eV (corresponding
to a speed of 2200 ms-1).
-8-
1.2
Lent 2014
Nuclear Reactions
A nuclear reaction is the process in which two nuclei, or one nucleus and a subatomic particle,
collide to produce one or more nuclides that are different from the nuclide(s) that began the
process. A nuclear reaction therefore always involves transmutation. Nuclear reactions may
involve alpha particles, neutrons, protons, electrons or positrons.
The first study of a nuclear reaction was carried out by Ernest Rutherford in 1919. He was the
first person to deliberately transmute one element into another: he used alpha radiation to
convert nitrogen into oxygen through the following reaction:
An alpha particle is absorbed and a proton is emitted. This reaction can also be written in the
compact form:
Here
is regarded as the target nuclide,
as the product nuclide. This experiment
showed Rutherford that hydrogen nuclei formed part of nitrogen nuclei (and therefore probably
other nuclei too) and led him to suggest that a hydrogen nucleus was possibly a fundamental
building block of all nuclei and perhaps also a new fundamental particle. He named it the
proton in 1920.
If a reaction like this is endothermic, then there is a threshold energy for it to be possible.
There is also the Coulomb barrier, which is the energy needed to overcome the electrostatic
repulsion of approaching positively charged nuclides.
Note that nuclear equations describe nuclear reactions without considering charges.
-9-
Lent 2014
Nuclear power generation relies on fission induced by incident neutrons. If a nucleus can
undergo fission regardless of the incident neutron energy (even if the probability of this
- 10 -
Lent 2014
occurring is low), the nucleus is referred to as fissile, whereas if there is a threshold energy
needed for fission to occur, the nucleus is referred to as fissionable. Examples of fissile nuclei
include
,
and
.
The most commonly used fissile nuclide in thermal reactors is
. When
absorbs a
neutron, it leads to the formation of the unstable radionuclide
, which immediately splits
into two smaller nuclei (typically of unequal mass), known as fission fragments. There are
many possible fission reactions, for example:
However, a neutron will not necessarily induce fission if it passes through the nucleus: for
example, fast neutrons are less likely to induce fission in
than thermal neutrons the
faster a neutron is travelling, the less time it spends inside the nucleus and therefore the less
opportunity it has to induce fission.
The fission event must emit more than one neutron if the reaction is to be sustained (and
therefore to create a chain reaction). Each fission of
generates an average of 2.4 neutrons,
and each fission of
gives an average of 2.9 neutrons. Each fission event typically
releases about 200 MeV of energy, about 35% of which is converted to electrical energy in
power stations.
The fission products typically exhibit a decay series. For example
, a fission product of the
above reaction, shows an isobaric (conserving mass number) decay series:
Neutron economy plays an important role in the design of nuclear reactors, since there are
several ways in which the released neutrons can be used up:
There is a certain minimum size of a chain reacting system, called the critical size, for which
the production of neutrons by fission just balances the loss of neutrons to the other processes,
and the reaction can be sustained independently. The associated mass is the critical mass. If
- 11 -
Lent 2014
more neutrons are lost than produced, the reactor is said to be subcritical, and if more are
produced than lost, the reactor is said to be supercritical.
Reactors can also be used to create (breed) fuel, the most common case being the production of
from fissionable
using fast (E = 1 20 MeV) neutrons:
- 12 -
Lent 2014
The neutron cross-section depends on the target type, the type of nuclear reaction (scattering,
fission etc) and the kinetic energy of the species involved. To a lesser extent, it also depends on
the angle between the incident neutron and the target nuclide and the target nuclide temperature.
Note that represents the microscopic cross-section and is a property of a given nuclide,
whereas the macroscopic cross-section, , takes into account the number of those nuclides
present. For mixtures of isotopes and elements, the macroscopic cross-sections add.
The range of a neutron is the distance it travels before being stopped and is a function of the
neutron energy as well as the capture cross-section of the material through which the neutron is
moving.
For fission to be induced by thermal neutrons, the binding energy of the thermal neutron to the
fissile nuclide must exceed the energy required for the nuclide to split (the fission barrier).
Heavy nuclides show very different cross-sections for induced fission, f, and for neutron
absorption followed by gamma emission, :
Nuclide
f (b)
(b)
3 x 10-6
737
4.8
7.5
530
48
6.8
6.0
99
6.5
5.7
2.7
4.8
5.8
269
6.5
5.0
586
3 x 10
752
-6
- 13 -
Lent 2014
1.2.2 Fusion
Fusion will not be discussed in detail in this course, but it is worth including at this point for
completeness.
Energy is given off when a nucleus becomes stable (i.e. when it approaches the maximum on
the graph in section 1.1.2). Whilst moving from heavier nuclei towards this maximum requires
the nucleus to split apart (fission), moving from lighter nuclei towards the maximum requires
two nuclei to combine and form a heavier one (fusion). The energy release per mass of nuclide
is much higher for fusion than fission.
The reaction of most interest for power generation, yielding 17.6 MeV per event, and with a
Coulomb barrier of only 0.68 MeV is the fusion of deuterium and tritium:
There are, however, many technical challenges and no commercial fusion reactors currently
exist. It is unlikely that any will be set up for some years, but fusion for power generation is a
prominent research topic. Experimental reactors are in the process of being built, such as ITER
(International Thermonuclear Experimental Reactor), which is planned to be completed by
2018.
- 14 -
1.3
Lent 2014
Nuclear Reactors
1.3.1 Components
The main components of nuclear reactors will be discussed in more detail in subsequent
sections, but a summary is given here.
The fuel can be in metallic, alloy or ceramic form and is generally contained within tubes made
of a metallic alloy (cladding). The most commonly used fuels are enriched uranium (uranium
in which the percentage of
has been increased), uranium oxides and plutonium oxides.
This cladding provides mechanical support to the fuel, prevents fission products from leaving
the fuel element and protects the fuel from corrosion caused by the coolant.
The fuel elements are typically arranged in a regular pattern (square, hexagonal etc) with the
moderator.
The moderator slows down neutrons to sustain the fission reaction with thermal neutrons.
The fuel-moderator assembly is surrounded by a reflector to direct neutrons towards the core
and to control neutron leakage (thereby improving neutron economy).
Outside, the reactor is surrounded by shielding that absorbs neutrons and gamma rays and
reduces the external radiation intensity to a tolerable level.
Control rods help to control the chain reaction by absorbing neutrons to maintain a steady
state of operation. Control rods are made from neutron-absorbing materials such as boron and
hafnium. There are usually two types of control rod in a nuclear reactor: rods for routine
control, which can be raised or lowered to increase or decrease the amount of heat being
generated, and safety rods, which can be lowered to shut the reactor down in an emergency.
The coolant removes the heat that is continually generated and is used to produce steam to
drive turbines for electricity generation. The coolant can be gas or liquid: examples include
light or heavy water, carbon dioxide, liquid metals and molten salts. A careful balance is
needed between the reduction in neutron economy due to the presence of the coolant and the
efficiency of heat removal.
- 15 -
Lent 2014
Mechanical properties the materials used should be strong enough to bear the normal
loads that the structure would be subjected to and withstand internal or external stresses,
and ductile enough to avoid any catastrophic failure.
Neutron properties for example, the fuel cladding materials need a low neutron
absorption cross-section, whereas the control materials need a high neutron absorption
cross-section.
Susceptibility to induced radioactivity as described above, absorption of neutrons can
lead to the formation of different isotopes, which may be radioactive. Considerations
include the abundance of the isotopes, their half-lives and the type of radiation that they
produce. For example, isotopes with a short half-life that emit low energy radiation are
much less of a concern than those with a long half-life that emit high energy radiation.
Radiation stability as we will see in later sections, radiation damage can lead to a
variety of effects, including the formation of voids, embrittlement, creep and hardening.
The requirements of materials used for different components within a nuclear reactor are
summarised in the table below.
- 16 -
Component
Cladding
material
Main requirements
Low neutron absorption
Stability under heat and radiation
Mechanical strength
Corrosion resistance
Good heat transfer properties
Lent 2014
Possible materials
Al, Be, Mg, Zr
Stainless steels
Ni-based superalloys
Refractory metals (Mo, Nb, Ti,
W etc)
Coolants
Shielding
material
Light water
Concrete
Most control materials
Metals (Fe, Pb, Bi, Ta, W, Boral)
- 17 -
Lent 2014
- 18 -
Lent 2014
The disadvantages of this type of reactor include limited efficiency and power capacity due to
relatively low maximum operating temperature (345C) this maximum was imposed due to
concern about reaction of CO2 with graphite at higher temperatures.
- 19 -
Lent 2014
- 20 -
Lent 2014
1.3.5.3 Advanced Gas-cooled Reactor (AGR): this type of reactor was developed from the
Magnox design, but operates at a higher gas temperature to improve thermal efficiency. It was
designed such that the final steam conditions are identical to those in conventional coal-fired
power stations, meaning that the same design of turbo-generator plant could be used. It uses
slightly enriched uranium as the fuel, graphite as the moderator and carbon dioxide as the
coolant.
It has a better thermal efficiency than a PWR, but the reactor core is larger for the same power
output and the fuel is used less efficiently, which counters the thermal efficiency advantage.
The AGR is designed to be refuelled without being shut down first, but fuel assembly vibration
problems during refueling at full power meant that refueling is now only performed at lower
power or when shut down.
- 21 -
Lent 2014
1.
2.
3.
4.
5.
Charge tubes
Control rods
Graphite moderator
Fuel assemblies
Concrete pressure vessel
and radiation shielding
6. Gas circulator
7. Water
8. Water circulator
9. Heat exchanger
10. Steam
Lent 2014
2. Radiation Damage
2.1
Introduction
Radiation damage refers to microscopic defects produced in materials due to irradiation, and
results in changes to their physical, chemical and mechanical properties. The interaction of
high energy subatomic particles and radiation with crystal lattices can give rise to a variety of
defects, including vacancies and self-interstitials. The majority of radiation damage is caused
by neutrons and fission fragments: other types of radiation generally have insufficient energy
(or are not produced in large enough quantities) to cause major damage.
Radiation damage has been widely studied because of its importance in structural components
in and near the cores of nuclear reactors. Such components are subjected to extreme conditions
of various kinds, including high temperature, high stress and corrosion, but neutron irradiation
leads to specific types of damage.
Note that the term radiation effects is generally used to refer to the effects on the behaviour
or properties of materials in the aftermath of radiation damage. Radiation effects will be
covered in more detail in subsequent sections.
2.2
The key point is that the incident radiation is sufficiently energetic to displace atoms from their
equilibrium sites in the crystal structure, since the binding energy of lattice atoms is typically
small compared to the energy of the impinging particles.
This creates vacancies and self-interstitials (one vacancy and one interstitial constitute a
Frenkel pair). The damage follows the sequence:
1. An energetic incident particle (usually a fast neutron) strikes an atom in the crystal
2. The transfer of kinetic energy to the atom in the primary recoil is large enough to
displace it from its lattice site and it becomes a primary knock-on atom, or PKA,
leaving behind a vacant site
3. The PKA moves through the lattice, creating further knock-on atoms in a displacement
cascade
4. The PKA and other knock-on atoms eventually come to rest as interstitial atoms
A displacement cascade (also known as a displacement spike) is illustrated in the image below.
There is a high density of vacancies in the core, with the surrounding material rich in
interstitials.
- 23 -
Lent 2014
The initial mean free path between collisions is of the order of 1 cm for fast neutrons but
decreases as energy is dissipated in successive collisions; the result is that displacement
cascades are concentrated in volumes 1 to 10 nm in diameter. Typical damage rates (measured
as displacements per atom per second, dpa s-1) range from 10-9 dpa s-1 in thermal reactors to
10-5 dpa s-1 in the first wall of proposed fusion reactors.
The displacement energy is the minimum energy that must be transferred to a lattice atom in
order for it to be displaced from its lattice site. Generally, an average displacement energy of
25 eV is assumed, but the actual value depends on a variety of factors including the material
and its crystallographic structure, the trajectories of the knock-on atoms, and the thermal
energy of the atoms. Generally, higher melting point metals tend to have higher displacement
energies, as illustrated below.
Variation of displacement energy as a function of melting temperature (from Murty & Charit, 2013)
- 24 -
Lent 2014
If the energy transferred by a knock-on atom to the struck lattice atom is less than the
displacement energy, the lattice atom will not be dislodged from its lattice site. It will instead
vibrate around an equilibrium position and the energy will be dissipated as heat.
Over the lifetime of a component, each atom could be displaced as many as 100 times. Clearly
such extreme conditions can have profound effects on the microstructure of the alloys involved.
These effects include dissolution of precipitates, changes in their morphology, and appearance
of non-equilibrium phases. In this course, we will focus on dislocations and formation of voids.
In principle, the vacancies and interstitials generated in displacement cascades might
recombine to restore the equilibrium structure. However, in practice, several factors can give a
substantial supersaturation of vacancies:
Self-interstitial atoms form stable clusters and are ultimately removed by dislocation
glide to other dislocations and grain boundaries; in effect, radiation damage has a bias
towards the production of vacancies.
Typically the vacancies and interstitials are created at a temperature high enough for
them to be mobile. The interstitials are more mobile, which is an additional factor
leaving the centres of displacement cascades vacancy-rich.
Dislocations, through the process known as climb, can act as sinks for vacancies and
interstitials, but the greater elastic strain around the latter again leads to their
preferential removal.
2.3
Dislocation loops
Dislocation loops can form at lower temperatures (T < 0.2 Tm, where Tm is the absolute melting
temperature of the irradiated alloy). The displacement cascade illustrated in section 2.2 can be
thought of as a core of vacancies surrounded by a shell of interstitials. If the vacancy core or
the interstitial shell collapse (condense) onto a close-packed plane, dislocation loops can be
generated: collapse of the vacancy core results in a vacancy loop, whilst collapse of the
interstitial shell results in an interstitial loop.
In ccp metals, the stacking sequence of close-packed planes (the {111} planes) can be
described by ABCABCABC. Both types of dislocation loop disrupt the stacking sequence of
the planes, resulting in a stacking fault: vacancy condensation produces an intrinsic fault and
interstitial condensation produces an extrinsic fault.
- 25 -
Lent 2014
Removal of a layer of atoms (i.e. following the formation of a vacancy loop) results in an
intrinsic (or single) fault in the stacking sequence, such as ABCAB/ABCABC (where /
indicates the missing plane of atoms). Insertion of an extra plane of atoms (following the
formation of an interstitial loop) produces an extrinsic (double) fault in the stacking sequence,
such as ABCAB/A/CABC.
The images below show an end-on view of a stack of close-packed planes in a crystalline metal:
(a) with a vacancy loop (intrinsic fault); (b) with an interstitial loop (extrinsic fault).
In these dislocation loops, the Burgers vector is normal to the planes and of magnitude equal to
the interplanar spacing (
). The dislocation is sessile (unable to glide).
The formation of vacancy-type and interstitial-type loops under irradiation is a major
contribution to the observed increase in dislocation density. The continuing evolution of
dislocation density under irradiation, however, involves many processes and is not susceptible
to analysis in terms of loop nucleation.
- 26 -
2.4
Lent 2014
Irradiation at higher temperatures (T > 0.2 Tm), leads to cavities or voids, the nucleation of
which is related to supersaturation not only of vacancies but also of dissolved helium atoms.
This helium arises from irradiation-induced transmutation reactions (of B, Ni and Fe)
accompanied by the emission of alpha particles: since alpha particles are positively charged,
they easily pick up electrons from the surrounding lattice and become elemental helium. The
rate of helium production is in the range of 0.512 atomic parts per million per dpa.
The precipitation of helium atoms results in the formation of small bubbles, which can
subsequently act as sinks for vacancies, thereby acting as a nucleation point for voids. The rate
of swelling is much greater than can be accounted for solely by the helium production rate, and
is mainly due to the condensation of vacancies.
Cavity nucleation under irradiation has the feature, not often encountered in metallurgical
precipitation, that the principal species (vacancies and helium atoms) are under continual
production; without this, cavities already produced would largely disappear on annealing.
From the first observations of cavity formation in 1967, it was recognised that this is a
particularly important form of radiation damage, leading to swelling and distortion of irradiated
components. At higher temperatures, it can cause hardening of irradiated alloys and associated
embrittlement and loss of ductility. The development of voids on grain boundaries from initial
helium bubbles also shortens creep life.
- 27 -
Lent 2014
3. Nuclear fuels
3.1
Introduction
The basic requirements of a nuclear fuel (apart from being easily fissionable and preferably
fissile) are:
good thermal conductivity (to allow heat generated during fission to be removed)
ideally a high melting temperature (to be able to run as hot as possible without melting)
mechanical stability
The design and operation of a reactor depend on the behaviour of the fuel. The relative cost of
the fuel in nuclear power generation is low, so it is possible to expend resources on optimising
it.
There must be enough fuel in a reactor for the chain reaction to be self-sustaining. The total
mass of fuel would be supercritical (leading to a runaway reaction) if it were in monolithic
form (i.e. a single block), but it is sub-divided to permit moderation and control, and to allow
heat extraction from the fuel to be efficient enough to prevent it from melting. The fuel can be
in the form of plates or rods (mostly rods), and it is isolated from the coolant by cladding (see
section 4).
Typically, in a PWR, the fuel rods are 45 m long, and the cladding is 912 mm in diameter
and 0.60.8 mm thick. Fuel assemblies have 200300 rods and there are 150250 such
assemblies in the core, giving approximately 80100 tonnes of uranium in the reactor.
Westinghouse fuel sub-assemblies for a PWR (ca. 1978) (from Frost, 1994; Ma, 1983)
- 28 -
Lent 2014
The linear heat generation rate (LHGR) is the heat generation rate (i.e. power) per unit length
of fuel rod (commonly expressed in kW/m) and is limited by the thermal conductivity of the
fuel and the need to avoid melting in the centre. Typical values are 1570 kW/m.
It is desirable to have the fuel rods inserted for longer in the reactor (so that they need to be
replaced less frequently) and to achieve greater burn-up (to allow more efficient use of the
fuel). Burn-up is a measure of how much energy is extracted from a primary nuclear fuel
source and is quoted as GW-days per tonne of U (or equivalent), GWd/t.
Generation II reactors were designed to achieve ~40 GWd/t. With newer fuel technology, and
particularly the use of nuclear poisons, these reactors are now capable of ~60 GWd/t.
Nuclear poisons (or neutron poisons) are neutron absorbers that are inserted into some reactors
to lower the high reactivity of the fresh fuel (the poisons are said to have negative reactivity): if
a reactor is designed to operate for a long period of time, more fuel than that needed for exact
criticality must be used, so neutron absorbers are added to control the reaction. The positive
reactivity from the excess fuel is balanced by the negative reactivity of the neutron absorber.
These nuclear poisons are often burnable, meaning that their effect wears off over the
lifetime of the fuel: ideally, the negative reactivity of the poison should decrease at the same
rate that the fuels excess positive reactivity is depleted. Examples include B4C-Al2O3,
borosilicate glass and Gd2O3.
Some more advanced designs are expected to achieve >90 GWd/t from higher-enriched fuel,
and eventually >200 GWd/t. Complete fission of all heavy metal in a breeder reactor, not just
fissile content but also any fissionable or fertile material, would yield ~1,000 GWd/t.
In a power station, high fuel burn-up is desirable for:
reducing the number of fresh nuclear fuel elements required and spent nuclear fuel
elements generated while producing a given amount of energy
reducing the potential for diversion of plutonium from spent fuel for use in nuclear
weapons
It is also desirable that burn-up should be as uniform as possible both within individual fuel
elements and from one element to another. Materials issues, particularly in the fuel and in the
cladding severely restrict the burn-up that can safely be achieved.
At present 98% of all nuclear-generated electricity comes from oxide fuels, but for historical
and scientific reasons we first consider metallic fuels.
- 29 -
3.2
Lent 2014
Metallic fuels
3.2.1 Uranium
Uranium is the basic nuclear fuel and is the basis for breeding new fuel. Pure uranium was used
in the earliest reactors (such as the Magnox reactors at Calder Hall). Metallic uranium and its
alloys are still used in teaching and research reactors for low-temperature operation, and are of
interest for some future designs of reactor.
In principle, metallic fuels have significant advantages:
Disadvantages include:
Irradiation stability and corrosion resistance can, however, be improved through alloying.
Uranium makes up about 4 ppm of the Earths crust (making it more common than elements
such as silver and mercury) and the amount of economically recoverable uranium in the world
has been estimated to be about 5.5 million tonnes.
Natural uranium has 0.7%
and 99.3%
(and 0.006%
). It is found in a variety of
minerals, such as pitchblende (U3O8), and uraninite (UO2). The main producers of uranium are
Kazakhstan (27%), Canada (20%) and Australia (20%), although a sizeable portion of uranium
is also produced by reprocessing spent fuel rods.
3.2.1.1 Structure
Uranium has three crystalline polymorphs:
Lent 2014
T-dependence of the lattice parameters of: (1) -U; (2) -U-15 at.% Pu (from Frost, 1994)
In polycrystalline samples, the anisotropic CTE causes problems when the temperature is
changed, since neighbouring grains (if unconstrained) would change shape in different ways on
heating or cooling. In a polycrystal, grains therefore exert stresses on each other and local
plastic flow, dislocation multiplication and hardening (and even failure of the material) can
occur.
If a polycrystal has grains in random orientations, no net shape change would be expected on
thermal cycling. However, rolled -U shows strong crystallographic texture (preferred grain
orientation), and textured polycrystals show thermal cycling growth. Note that growth in
this context means shape change at constant volume. The growth arises from relative
movement between neighbouring grains in different orientations combined with stress
relaxation in some of the grains by plastic deformation or creep.
The images below show the effect of thermal cycling growth in highly oriented fine-grained
-uranium between 50C and 500C for 1300 cycles (top) and 3000 cycles (bottom).
- 31 -
Lent 2014
Due to its greater symmetry, the gamma phase of uranium does not exhibit thermal cycling
growth, so gamma stabilising alloying additions such as Al, Mo and Mg can help avoid this
effect, as shown below. U-Mo alloys typically contain at least 6 wt% Mo in order to avoid
thermal cycling growth.
Lent 2014
It is worth noting that the mechanical properties depend on texture, fabrication history and heat
treatment. The tensile properties are also affected by impurities such as carbon, fission
products or alloying elements, and the strength also decreases dramatically with increasing
temperature.
- 33 -
Lent 2014
3.2.1.5 Corrosion
Uranium reacts rapidly with most environments (air, oxygen, hydrogen, water etc). The UO2
surface layer that forms is not quite protective: at high temperatures, as the film thickens, it
cracks and crumbles, exposing fresh uranium underneath. Similarly, in water, the UO2 film
provides reasonable corrosion resistance at low temperatures (50-70C), but at higher
temperatures, the oxide becomes porous and the protection is lost.
Irradiation enhances corrosion, but this will be discussed in more detail in section 7.
3.2.1.6 Irradiation growth
Irradiation growth is a form of dimensional instability that occurs under irradiation without the
need for an applied stress, at relatively low temperatures (~300C). As in thermal cycling
growth, the volume of the material remains constant during irradiation growth (essentially,
material is moved from one place to another), so this is different to radiation swelling (see
section 3.2.1.7).
Under a neutron flux, even single crystals of -U show irradiation growth (although note that
-U does not, due to its isotropic nature). The growth is accompanied by hardening and
embrittlement, and arises from the generation of dislocation loops:
Interstitial and vacancy loops form on different planes because of the thermal spike and the
anisotropic CTE.
Consider the effect of heating an individual grain relative to its surroundings: it experiences
tension parallel to [010] (because the CTE in that direction is negative). Interstitial atoms will
then preferentially condense on (010), thereby giving an expansion to relieve the tensile stress
(analagous effects occur for vacancy loops on other planes). There is therefore a net expansion
parallel to [010].
If the material is polycrystalline, it will not show irradiation growth if there is no
crystallographic texture (individual grains would change shape but there would be no net shape
change), but irradiation growth will occur if it has preferred orientation.
- 34 -
Lent 2014
Swelling rate of uranium phases in the burn-up range 0.20.5 at.% (from Frost (1994))
- 35 -
Lent 2014
As illustrated below, voids in -U have a characteristic faceted shape, whilst those in cubic -U
are more rounded (due to the more isotropic nature of -U).
Pores in irradiated -U10Zr (wt%) (left) and in -U10Zr (right) (from Frost (1994))
- 36 -
Lent 2014
- 37 -
Lent 2014
3.2.2 Plutonium
Plutonium can also be used as a nuclear fuel in nuclear reactors and in space applications.
is the major fissile isotope of plutonium and has a high fission cross-section. Plutonium
is only found naturally in trace quantities and is mainly produced artificially through the
transmutation of
.
It can be recovered from spent fuel in thermal reactors, and depleted uranium can be kept
together with plutonium for fuels used in fast breeder reactors. Separated plutonium can also be
used in plutonium-burning reactors.
There are 6 phases of plutonium, which are only stable in limited temperature ranges. It is
monoclinic at room temperature, but also has orthorhombic, fcc, bct and bcc phases, and it also
readily undergoes martensitic transformations. It has a relatively low melting point of 640C.
Plutonium has a lower critical mass than uranium, and is also toxic and pyrophoric as well as
being very sensitive to corrosion. It is therefore difficult to work with, but a variety of
fabrication techniques can be used. Its mechanical properties depend on its phase (some phases
are very brittle, but others are relatively ductile), but also on impurity and defect concentrations.
Its properties therefore do not allow it to be used in pure form, so a variety of alloying
additions (Al, Ga, Mo, Th, Zr etc) are added. Zr is often added to increase the melting
temperature and to reduce interdiffusion with the stainless steel cladding. However, the
temperature gradient in alloy fuel rods leads to composition variation, driven by the Soret
effect (different atoms exhibit different responses to a temperature gradient).
- 38 -
Lent 2014
3.2.3 Thorium
Thorium is another nuclear fuel that has not yet been used to its full potential.
is a fertile
isotope that could produce fissile
upon capturing a neutron, so is therefore an important
breeder material.
is the only naturally-occurring isotope and has a half-life of 14 billion
years (!).
Thorium is far more abundant than uranium in nature: most rocks and sands contain minute
amounts of thorium, and monazite is a rare earth phosphate mineral, containing 67 wt%
thorium.
Thorium has two phases, both of which are cubic, and it has a melting point of ~1750C. Its
mechanical properties are sensitive to impurities, cold work and crystallographic texture, and it
is sensitive to corrosion (although much less so than plutonium). Its mechanical properties and
corrosion resistance can be improved by alloying.
The commercial use of thorium has a number of obstacles to overcome: a large amount of
R&D and testing are still needed and the costs of fuel fabrication and reprocessing are
high.
also becomes contaminated with
, which decays to daughter nuclides that are
high-energy gamma-emitters (and therefore difficult to handle), such as
.
- 39 -
3.3
Lent 2014
Ceramic fuels
The essential requirements remain that there should be a sufficient number density of fissile
atoms (and we would like to avoid high levels of U enrichment), and that the other nuclides
(non-fissile components) should have a low mass number and small neutron absorption
coefficient.
Potential advantages of ceramics:
higher fuel and plant operating temperature (higher melting temperature than metals)
excellent corrosion resistance
low thermal expansion coefficients
Disadvantages of ceramics:
- 40 -
Lent 2014
3.3.1 UO2
UO2 is a precursor to U metal and was first used as a blanket fuel, resistant to high-temperature
water (a blanket fuel is a layer of material containing fertile isotopes that is placed around the
reactor core as a reflector or absorber, but which is also used to breed additional fissionable
material). However, it is now very well studied, and has been in reliable use as a nuclear fuel
for nearly 50 years.
In UO2 fuel, either natural or slightly enriched uranium (0.7 4%
) can be used. If enriched
uranium is used, it is good practice to vary the enrichment across the core, as illustrated below.
This is because irradiation is usually non-uniform and tends to be higher in the centre of a fuel
rod, so having a varying enrichment can allow burn-up to occur more evenly.
Plan of the fuel sub-assembly and reactor core for a PWR (from Frost, 1994)
Note that Fast Breeder Reactors use a mixed oxide, MOX: (U0.75Pu0.25)O2. The oxides UO2 and
PuO2 show complete mutual solubility:
Solidus and liquidus lines in the UO2-PuO2 equilibrium phase diagram (from Frost, 1994)
- 41 -
Lent 2014
A high density is desirable to ensure a high density of fissile or fissionable atoms, and good
thermal conductivity. However, some residual porosity is useful to retain fission product gases
and to allow densification on heating to offset irradiation swelling.
UO2 can be processed by conventional ceramic powder sintering into bulk shapes such as
pellets, rods and tubes with densities typically 9397% of the theoretical density (a lower
density of 85% is used for Fast Breeder Reactors, to allow greater burn-up). Pellets are usually
used (see image below), since it is difficult to manufacture a whole fuel rod from a ceramic.
Sintering must be performed in an inert (or reducing) atmosphere, because sintering in air can
lead to the formation of other uranium oxides such as U3O8 the different densities of the
various oxide phases causes problems during sintering.
3.3.1.1 Structure
UO2 has the CaF2 structure (fcc, a = 5.47 ): the calcium ions occupy the face-centred lattice
sites whilst the oxygen anions sit in tetrahedral interstices. The large octahedral interstice in the
centre of the unit cell is empty and can therefore accommodate fission products. This is
important for radiation stability (since it prevents the products from diffusing elsewhere and
forming voids).
- 42 -
Lent 2014
UO2 shows no polymorphic phase changes on heating and its melting temperature (2865C) is
much higher than that of uranium metal. UO2 does restructure by grain growth and void
migration, especially above 1700C.
It has a low tensile strength of approximately 35 MPa and a Youngs modulus of
approximately 170 GPa. It rapidly loses strength and becomes ductile above 11001400C (the
transition temperature depending on strain rate and grain size), but exhibits cracking on reactor
start-up and shut-down. This cracking occurs primarily due to thermal stresses rather than due
to irradiation effects.
The image below shows crack distribution in a fuel pellet, with a superposed strength vs
temperature curve for (U, Pu)O2. The pellet is ductile in the centre, where the temperature is
highest, but becomes brittle closer to the edges, where the temperature is lower.
The amount of gas released depends on a variety of factors, including porosity, irradiation time
and irradiation temperature. The centre temperature of the rod in thermal reactors is typically
restricted to ~1400C (note that in FBRs the temperature may initially be higher to promote gas
release.)
Irradiation swelling of UO2 is low, as illustrated below.
- 43 -
Lent 2014
Irradiation swelling: (a) in U metal in a Magnox reactor; (b) in UO2 in a PWR (from Frost, 1994)
Failure of oxide fuel rods occurs mainly due to pellet-cladding mechanical interaction (PCI).
This can be alleviated by:
leaving sufficient space between the fuel and cladding (but note that this can affect the
efficiency of heat transfer)
using graphite as a liner to provide lubrication
using smaller diameter rods at lower LHGR
UO2 can deviate significantly from stoichiometry. This is an important issue affecting
properties and possible burn-up. As fuel is burned, the oxygen:metal ratio increases (because
not quite all the oxygen released can combine with fission products). It is good practice to start
with hypostoichiometric fuel (where oxygen:metal 1.97) as oxygen release is detrimental
for the cladding.
- 44 -
Lent 2014
In the ceramic UO2, transport (i.e. diffusion) rates are low (compared to those in metallic U),
but the movement of grain boundaries during grain growth sweeps gas from grain interiors to
grain boundaries, where the bubbles then coarsen by diffusion of gas atoms along the grain
boundaries, as shown below.
Gas bubbles on grain boundaries in UO2 tested in an AGR (from Frost, 1994)
Bubbles start from residual porosity in the sintered pellets and then migrate up the temperature
gradient in the fuel rod. UO3 is the main volatile species.
The migration of bubbles up the temperature gradient
generates a void along the central axis of the fuel pellet.
This image shows thermal restructuring in a 7.5 mm
diameter 94% dense fuel pellet of (U, Pu)O2, tested in a
FBR at a LHGR of 45 kW/m and a burn-up of 40 GWd/t.
Grain growth, cracking and a central void can be seen (from
Frost, 1994).
In MOX fuel rods, similar processes transport U from the centre to the surface, leaving the
centre Pu-rich.
- 45 -
Lent 2014
They are widely recognised as intrinsically better fuels for the future, in particular as advanced
fuels for FBRs. They would permit moderate LHGR (~70 kW/m), high burn-up (>150 GWd/t),
and efficient operation as they have a higher density of metallic atoms than UO2.
UC has been used in thermal and FBRs, and has been more extensively studied than UN, but
UN is of particular interest for easier reprocessing (important for the FBR fuel cycle).
The carbide and nitride fuels show similar irradiation effects to UO2, but they have greater
stability and better thermal conductivity, thus permitting higher LHGR and greater burn-up.
- 46 -
Lent 2014
4. Cladding
4.1
Introduction
isolate the fuel from the coolant (this requires chemical stability and mechanical
integrity)
have a low neutron absorption cross-section (to allow neutrons to pass through the
cladding without being absorbed)
be corrosion-resistant, and in general be compatible with the fuel and coolant
conduct heat well (to allow heat to be transferred from the fuel, through the cladding, to
the coolant: heat transfer from fuel to cladding should also be considered, as well as
heat transfer through the cladding)
have a melting point well above the operating temperature of the reactor (note that new
designs of reactor may well have higher operating temperatures)
have adequate strength and ductility, especially to withstand swelling of the fuel
have high stability under irradiation
have a low induced radioactivity (to facilitate recycling after use)
The cladding is one of the most critical components in a reactor, and its performance
(especially irradiation growth and corrosion resistance) is often the key factor limiting
attainable burn-up.
The most commonly used cladding materials are stainless steels (for more extreme conditions
in FBRs) and zirconium alloys (in thermal reactors).
4.2
Austenitic (or ) stainless steels have an fcc structure. In low carbon steels, austenite is only
stable above the eutectoid temperature (730C), but different alloying additions can stabilise
austenite to lower temperatures.
Austenitic stainless steels are the cladding materials of choice in FBRs where operating
conditions are particularly severe:
- 47 -
Lent 2014
The most commonly used austenitic stainless steels are 304 and 316, typical compositions of
which are tabulated below.
Fe
Cr
Ni
17.5-20%
8-11%
16-18.5%
10-14%
Mo
2-3%
Mn
Si
<2% <1%
<0.045%
<0.03%
<2% <1%
<0.045%
<0.03%
Under irradiation, austenitic stainless steels suffer from phase instability, void swelling,
irradiation creep, and irradiation-induced changes in mechanical properties, all of which are
linked to each other. However, it is void swelling that ultimately determines the creep
behaviour and other mechanical properties.
In both cases, a nickel nucleus absorbs a neutron, generates an alpha particle (which easily
gains electrons to become elemental helium) and undergoes transmutation to form an iron
nucleus. The figure below shows cross-sections for (n, ) reactions as a function of neutron
energy for the most common elements in stainless steels: it can be seen that Ni has a much
larger cross-section than any of the other elements.
Cross-sections for (n, ) reactions as a function of neutron energy for common elements in stainless
steels (from Frost, 1994)
- 48 -
Lent 2014
Due to their small size, helium atoms can easily diffuse interstitially and form bubbles.
However, voids that form under irradiation are much larger than can be explained only by the
production of helium: as described in section 2, helium bubbles act as sinks for vacancies and
therefore act as nucleation sites for voids.
Austenitic stainless steels show good dimensional stability under irradiation, until the onset of
void swelling. The figure below schematically shows void swelling behaviour as a function of
radiation dose, illustrating the different stages of void swelling. The initial incubation period
represents the neutron dose required to produce sufficient helium to make void nucleation
possible: this incubation period depends on a variety of factors, including temperature,
composition and microstructure.
After the incubation and transient period (which is highly variable and depends on many
factors linked to the distribution of solute elements), the swelling enters steady state. The
steady state swelling rate in austenitic stainless steels is typically ~1% per dpa (in comparison,
the rate in ferritic steels is 0.1% per dpa), although the rate is lower at both high and low
temperatures. This is illustrated in the figure below, which schematically shows the three
regimes of steady-state swelling rate in 316 stainless steel.
- 49 -
Lent 2014
In region I (at low temperatures) direct recombination of vacancies and interstitials can occur
(due to lower mobilities at lower temperatures): the swelling rate is therefore low but increases
with increasing temperature (as mobility increases).
In region II, the steady state swelling rate is high, and can only be explained by an excess of
vacancies.
In region III (at higher temperatures), the swelling rate decreases with increasing temperature,
since the rate of vacancy emission from voids is greater than the flux of vacancies joining the
voids.
Under neutron irradiation the maximum total swelling reported for an austenitic stainless steel
is 88% (at 510C), although the largest reported swelling is 260% for irradiation by 140 keV
protons at 625C!
The images below show void swelling at different temperatures in 316 stainless steel at
~1.41027 neutrons/m2 (En > 0.1 MeV), equivalent to ~70 dpa (from Frost, 1994).
- 50 -
Lent 2014
The image below shows swelling (~10% linear, 33% by volume) of a 316 stainless steel
cladding tube (not containing fuel) irradiated at 1.51027 neutrons/m2 (En > 0.1 MeV),
equivalent to ~75 dpa, at 510C. Note that all relative proportions are preserved during
swelling (from Frost, 1994).
The swelling of the cladding can be quite variable, as illustrated in the left-hand image (a)
below. This shows the top of a bundle of fuel rods irradiated to 2.1x1027 n/m2 (E > 0.1 MeV):
the rods vary in length as a function of gradients in dose, non-uniformity of composition and
variations in crystallographic texture (as a result of processing).
In comparison, the right-hand image (b) shows an undistorted fuel-rod assembly manufactured
using a non-swelling cladding (see below) that has been irradiated at 1.9x1027 n/m2 (E > 0.1
MeV) (from Frost, 1994).
- 51 -
Lent 2014
- 52 -
Lent 2014
The more mobile Cr and Fe atoms diffuse away from voids to such an extent that the matrix
away from the voids converts to phase ferrite (bcc) and only the regions close to the voids
retain the original austenite phase. This occurs because nickel (which is itself fcc) is an
austenite stabiliser whereas chromium (which is bcc) is a ferrite stabiliser.
- 53 -
Lent 2014
The figure below shows dislocation density as a function of irradiation dose in annealed and
cold-worked austenitic stainless steels, irradiated at 500C (from Frost, 1994).
The irradiation drives the dislocation density to the same steady-state value, independent of the
starting value.
This steady-state arises due to a balance between irradiation causing an increase in dislocation
density (as discussed in section 2), whilst also causing an increase in mobility (due to heating
of the material and therefore annealing), which enables dislocations to be removed and
therefore decreases the dislocation density: at steady-state, dislocations are generated at the
same rate as they are removed, meaning that the dislocation density reaches a constant value.
In an annealed sample, the initial dislocation density is low, so irradiation leads to an overall
increase in dislocation density. However, in a cold-worked sample, the initial dislocation
density is already high, so the effect of increased mobility means that dislocations are removed
faster than they are created (until steady state is reached), so that there is an overall decrease in
dislocation density.
The absolute value of the steady-state depends on a variety of factors including temperature
and neutron energy. The higher the temperature, the more effective the annealing-out of
damage, so the lower the steady state dislocation density.
- 54 -
4.2.4 Creep
Irradiation accelerates creep by orders of
magnitude, as seen in this graph, which shows the
influence of neutron irradiation on creep rate of an
annealed steel (from Frost, 1994). Components can
be readily designed to avoid normal thermal creep,
so irradiation creep is the main issue. There is a
complex interaction of displacement rate and
temperature, and it is difficult to do in-reactor
experiments.
- 55 -
Lent 2014
Conversely,
irradiation
at
high
temperatures can lead to softening as a
lower steady-state dislocation density is
reached (see also figure in section
4.2.3).
This graph shows the evolution of yield
strength of 316 stainless steel irradiated
at temperatures where there is void
swelling (from Frost, 1994).
- 56 -
Lent 2014
Lent 2014
Samples that have been tensile tested show shearing of voids and localised deformation. The
image below shows an annealed 304L stainless steel after irradiation to 38 dpa at 387C
followed by tensile testing at 370C (from Frost, 1994)
- 57 -
Lent 2014
4.2.6 Summary
The effect of irradiation on the properties of austenitic stainless steel:
- 58 -
4.3
Lent 2014
Ferritic alloys
Ferritic () stainless steels have the bcc structure. Although they exhibit a substantially lower
steady-state void swelling than austenitic stainless steels, they have not been the materials of
choice for cladding, partly because they show a ductile-to-brittle transition at low temperatures,
which is made worse by irradiation.
The figure below shows the effects of irradiation on the fracture energy and ductile-brittle
transition temperature in a base metal (left) and the heat-affected zone (HAZ) of welds in this
metal (right) (from Frost, 1994).
- 59 -
4.4
Lent 2014
Zirconium alloys
Zirconium has a very low thermal neutron absorption cross-section (30 times lower than Fe)
and the resulting neutron efficiency is the main reason for its use. Zr alloys are unusual in
engineering applications, though adequate ductility and good corrosion resistance have led to
widespread use in the chemical industry.
The alloying used to improve the corrosion resistance of Zr was discovered through accidental
contamination by stainless steel, which led to the development of the Zircaloys.
Zircaloy 2 is used in BWRs. A later alloy, Zircaloy 4 contains no Ni (to reduce hydrogen pickup) and is used in PWRs. Zr-Nb alloys are used in CANDU reactors. The compositions of
these alloys are tabulated below.
The Zr alloys are used in these thermal reactors mainly as fuel cladding, but also for some
other structural components. These alloys remain under active research to develop processing
conditions and compositions modified within existing specifications to improve their corrosion
resistance and texture.
Zircaloy-2
Zircaloy-4
Zr-Nb
Sn
1.21.7
1.21.7
Fe
0.070.2
0.180.24
Cr
0.050.15
0.070.13
Ni
0.030.08
Nb
2.42.8
10001400 ppm
0.090.13
Lent 2014
4.4.1 Structure
Zirconium has a hexagonal phase, which transforms to a bcc phase at 865C (which
therefore limits the temperature range of use). The melting temperature is 1860C.
The Zr alloys are used in the phase condition, with a microstructure governed by the
martensitic or bainitic transformation on cooling from .
The phase is highly anisotropic:
Lattice parameter
Youngs modulus, E
CTE
In basal plane
a = 3.23
c = 5.15
99 GPa
125 GPa
5.20 10
-6
1.04 10-5
Unlike most hexagonal materials, it does not slip on the basal (001) planes, but instead slips on
prismatic planes.
At higher temperatures, there are additional slip systems and twinning can also occur. This
gives the equivalent of the five independent deformation systems needed for general plastic
flow, so zirconium (unlike many hexagonal materials, which do not have sufficient slip
systems) is relatively ductile.
Zircaloys are processed by hot rolling in the phase at ~1050C, which gives strong
crystallographic texture (preferred orientation).
- 61 -
Lent 2014
Unlike stainless steels, there is no significant void formation (since many gases are soluble in
Zr), but cavities can form in the interstitial-free zone near grain boundaries.
Under irradiation, precipitates dissolve, decompose and reprecipitate, which affects corrosion
resistance.
- 62 -
Lent 2014
- 63 -
Lent 2014
4.4.2.4 Corrosion
Zirconium has a good corrosion resistance under normal service conditions. However, it reacts
readily with water at higher temperatures, leading to the formation of hydrogen gas. This
reaction led to the Fukushima nuclear accident in 2011: the reactors had shut down following
an earthquake, but during the subsequent tsunami, the plant flooded, cutting power to the
cooling system and causing the reactors to overheat .
The corrosion of zirconium is also catalysed by Ni (which is present in Zircaloy-2), and
Zircaloy-4 (which does not contain Ni) was developed to avoid this.
The cladding also suffers from internal attack: this micrograph shows stress corrosion cracking
(see section 7) of recrystallised zirconium in the presence of the fission product iodine.
- 64 -
Lent 2014
5. Moderators
5.1
Introduction
Moderators are needed in thermal reactors to slow the fast neutrons (with mean energy
~2 MeV) into the thermal energy range (~0.025 eV), without absorbing so many neutrons that
the chain reaction cannot be sustained. Moderators therefore need a high scattering crosssection, s, and a low absorption cross-section, a, such as would be found in the top left of the
selection map below:
The mass of the target nuclide is important as energy transfer to it is most efficient when its
mass is closest to that of the neutron.
The factor by which the neutron energy is reduced per collision, , is independent of the
neutron energy and is given by:
- 65 -
Lent 2014
- 66 -
5.2
Lent 2014
Graphite
Graphite is used as a moderator in AGRs, VHTRs and is proposed as a protective liner for the
first wall in fusion reactors. It is also used as a coating inside cladding to reduce friction with
the fuel. Existing reactor designs using graphite may be in service for more than 30 years.
Graphites neutron absorption is so low that there may be little or no need to enrich the U fuel
(and it is desirable to avoid the cost of enrichment).
5.2.1 Structure
Graphite is hexagonal (c = 6.71 , a = 2.46 ). It is used in
bulk polycrystalline form, with grain size typically Lc 60
nm, La 80nm, processed to be as isotropic as possible to
give better stability under irradiation.
Under irradiation, the c lattice parameter increases and the a
parameter decreases. Macroscopic dimensions parallel to the
same axes change at least as much and these changes mostly
occur at constant volume.
The origin of the growth is interstitial dislocation loops, and the basal plane contraction reflects
the collapse of vacancy loops. This graph shows irradiation-induced dimensional changes in
pyrolytic graphite (from Frost, 1994).
- 67 -
Lent 2014
The graphs below show the variation of principal CTEs of highly oriented pyrolytic graphite as
a function of neutron dose, at various irradiation temperatures (from Frost, 1994).
Additionally:
the thermal resistance doubles
strength, Youngs modulus and creep rate all increase under irradiation
the reactivity is increased (which is of most interest for oxidation)
- 68 -
Lent 2014
The Wigner energy is well characterised, but poorly understood and is still under active
investigation.
- 69 -
5.3
Lent 2014
5.3.1 ZrH
In hydrides such as ZrH, the density of H atoms can be higher than in liquid H2. Such materials
are attractive for moderators if it is desirable to minimise core weight and volume.
Examples of applications are in power plants for polar regions, space systems and other planets.
Zirconium hydrides have also been used in the TRIGA reactor, which is a reactor designed
for use by scientific institutions and universities for purposes such as teaching, private
commercial research, non-destructive testing and isotope production.
There are many hydrides of Zr. They are very brittle, but appear to be dimensionally stable
under irradiation. They also have quite poor thermal conductivity and therefore require
additional cooling in reactors operating at high to moderate temperature.
5.3.2 Beryllium
Beryllium has been used in some reactors, and beryllium oxide is proposed as a moderator for
high-temperature gas-cooled reactors. It has many excellent characteristics, such as low
absorption cross-section (superior to graphite in this respect) and high melting temperature.
However, it has poor ductility, is very toxic and very expensive compared to graphite.
5.4
Liquid moderators
Liquid moderators have the advantage that they also serve as coolants. The most common
liquid moderator is H2O (light water), but D2O (heavy water) is used in CANDU reactors.
Typical conditions:
The moderating power of light water is excellent, but its neutron absorption cross-section is
such that it necessary to use enriched uranium (2-3%), which therefore incurs the significant
extra cost of enrichment.
The advantage of heavy water is that its absorption cross-section is so low that enrichment is
not required.
However, there is a problem that radiolysis of water generates many active species that can
lead to stress-corrosion cracking of stainless-steel reactor components (see section 7).
- 70 -
Lent 2014
6. Control rods
6.1 Introduction
As discussed in section 1.3, control rods help to control the chain reaction by absorbing
neutrons to maintain a steady state of operation.
There are usually two types of control rod in a nuclear reactor:
1. Rods for routine control, which can be raised or lowered to increase or decrease the
amount of heat being generated
2. Safety rods, which can be lowered to shut the reactor down in an emergency.
The requirements of control rod materials are:
The choice of materials is also influenced by the energy of neutrons in the reactor. The
selection map below shows neutron absorption cross-sections for thermal neutrons, in which
materials suitable for use in control rods are highlighted:
- 71 -
Lent 2014
Control rods often have the form of stainless steel tubes filled with neutron absorbing pellets or
powder. Swelling of the control material can cause deformation of the rod (similar to the
interaction between fuel and cladding), leading to its premature replacement.
Control rods are typically about the same diameter as fuel rods, and are distributed evenly
within the fuel assembly: a single large control rod would lead to non-uniform temperature and
fuel fission.
However, it is very expensive and not readily available, which limits its use in civilian reactors,
though it is used in some US Navy reactors.
Dysprosium titanate is a material currently undergoing evaluation for PWR control rods as a
potential replacement for Ag-In-Cd alloys. Advantages include a much higher melting point,
and the fact that it does not swell or react with cladding materials or produce radioactive waste.
- 72 -
Lent 2014
Introduction
There is a wide range of corrosion problems in nuclear reactors, associated with different
components and aggressive corrosion environments. Our focus is on those problems
exacerbated by irradiation, notably irradiation-assisted stress-corrosion cracking (IASCC). This
is known to occur very widely and is a major factor limiting component lifetime.
The images below show irradiation-assisted stress-corrosion cracking in a cold-worked
austenitic 316 stainless steel bolt. The crack location received a neutron dose corresponding to
~7 dpa at ~310C.
7.2
Stress-corrosion cracking involves cracking of a material by the combined action of stress and
a corrosive environment, when no corrosion or cracking would be expected if only one of these
factors was operating. The stresses may be due to applied loads, residual stresses from the
manufacturing process, or a combination of both.
SCC is found in alloys, not ultra-pure metals, and susceptibility to it is highly dependent on
alloy composition. SCC typically only occurs in a specific environment for a given alloy.
The cracking itself can be transgranular (as shown in (a) below, TG) or intergranular (as shown
in (b), IG).
- 73 -
Lent 2014
Transgranular cracking is quite rare: in most cases, the cracking is intergranular, and a typical
fracture surface is shown below.
http://corrosion.ksc.nasa.gov/stresscor.htm
Cross sections of SCC frequently show branched cracks and this river branching pattern is
unique to SCC.
http://corrosion.ksc.nasa.gov/stresscor.htm
- 74 -
Lent 2014
Interestingly, alloys that are intrinsically corrosion-resistant tend to be susceptible to SCC and
there are cases where the corrosion rate of an alloy is essentially zero, yet it is highly
susceptible to transgranular SCC. For example, in Fe-18Cr-xNi alloys, SCC susceptibility is
greatest for 8wt%Ni, the composition that is usually chosen because it exhibits the lowest rate
of general corrosion.
Cracking typically occurs in two stages: initiation and propagation. Following initiation, cracks
propagate at a slow rate until the stresses in the remaining material exceed the fracture stress
and failure occurs. Some alloys that are immune to crack initiation in a particular environment
show susceptibility to crack propagation when they are pre-cracked.
Cracks in SCC would not advance without removal of metal at the crack tip. The crack growth
mechanism that is thought to be involved is anodic dissolution at the crack tip. The current
density at the crack tip must significantly exceed that from the crack walls because otherwise
the crack would blunt.
The corrosion resistance of alloys is usually largely because of a passive oxide film on their
surface. Given that corrosion-resistant alloys are particularly susceptible to SCC, disruption of
the passive layer is a likely cause.
As illustrated, below, plastic deformation (slip) in the metal reveals fresh metal that suffers fast
corrosion before repassivation occurs. Intermediate rates of repassivation are most dangerous,
because fast repassivation does not allow any significant corrosive attack whilst slow
repassivation allows corrosion to blunt the crack.
A given degree of plastic flow can occur in different ways. In some alloys (those with high
stacking fault energy), dislocations can easily cross from one slip plane to another. In this way
the deformation is rather diffuse. When the stacking fault energy is low, cross-slip is difficult;
in this case slip tends to be localised on planes that have active dislocation sources or relatively
low barriers to dislocation glide.
- 75 -
Lent 2014
The key point is that, in the latter case, slip is more localised, and the slip steps are larger,
promoting SCC. In austenitic stainless steels, nickel raises the stacking fault energy, causing
slip to be more diffuse, and the time to failure is increased.
The image below illustrates the progression of SCC:
- 76 -
7.3
Lent 2014
It is clear that neutron irradiation has an effect on SCC, but the exact mechanism by which it
occurs is not known. The threshold fluence beyond which irradiation effects become dominant
in BWRs corresponds to 0.3-0.7 dpa.
The graph below shows the effect of neutron dose on SCC (note that IGSCC is intergranular
stress-corrosion cracking).
Inevitably, there are many possible influences of irradiation and existing theories to explain
IASCC, including: radiation-induced segregation, radiation hardening, localised deformation,
selective internal oxidation and irradiation creep. This schematic shows mechanisms of SCC in
austenitic stainless steels that might be irradiation-assisted.
- 77 -
Lent 2014
OH + OH H2O2
H + O2 HO2
These reactions raise the electrochemical corrosion potential, and can significantly increase
corrosion rates.
This can be effectively countered by addition of H2 to the water, which rapidly reacts with the
oxidants: addition of only 500 ppb is sufficient to suppress the effects of radiolytic formation
of oxidants.
However, while this works in PWRs, it fails in BWRs. This is because in BWRs, water rises
along the fuel rods, and about 25% of the way up steam starts to form. Any H2 added to the
water also boils off (i.e. it is partitioned preferentially to the steam phase) and therefore cannot
remove oxidants from the water.
Lent 2014
graph below, which shows irradiation creep in a bolt torqued to 250 MPa and then irradiated
with neutrons. The effect of the irradiation is stress relief.
However, on the other hand, irradiation will ultimately generate stresses (particularly through
swelling), which will enhance SCC.
7.3.2.2 Segregation
For good corrosion resistance, the composition should be uniform to avoid a variation in
electrochemical potential. However, the composition of alloys tends to deviate from bulk
values at grain boundaries due to segregation and precipitate formation on the boundaries.
Austenitic stainless steels rely on chromium for corrosion resistance, but the formation of Crrich precipitates at their grain boundaries draws Cr from the neighbouring matrix, leaving it
more susceptible to corrosion. Such effects are known to underlie the intergranular SCC of
stainless steels in hot water and these effects can be exacerbated under irradiation (although the
correlation is noisy).
Another irradiation effect observed in austenitic stainless steels cladding tubes is the inverse
Kirkendall effect (section 4.2.2), in which the more mobile Cr and Fe atoms diffuse away from
voids to such an extent that the matrix away from the voids converts to ferrite and only the
regions close to the voids retain the original austenite phase. Ferrite is more susceptible to
corrosion than austenite, meaning that corrosion can be enhanced by the phase transformation.
- 79 -
Lent 2014
7.3.2.3 Hardening
Hardening can enhance SCC, but if the same degree of hardening is induced in a material by
irradiation or by cold-work, it is the irradiated sample that shows greater enhancement of SCC.
Irradiation can lead to the localisation of strain into dislocation channels. This localisation
within grains puts particular pressure on grain boundaries and can cause grain-boundary sliding.
This is particularly effective in rupturing the passive oxide film and promoting SCC.
- 80 -
- 81 -
Lent 2014
Lent 2014
Glossary
Binding energy
Burn-up
Cladding
Control rods
Coolant
Critical mass/size
Cross-section
Dislocation loops
Displacement cascade/spike
Displacement energy
Fertile
Fissile
- 82 -
Lent 2014
Fluence
Fuel element
Half-life
Irradiation growth
Irradiation swelling
Isotopes
The heat generation rate per unit length of fuel rod (commonly
expressed in kW/m)
Moderator
Nuclear poison
Radiation damage
Radiation effects
Transmutation
Zircaloys
- 83 -
Abbreviations
AGR
BWR
CTE
dpa
FBR
GCR
GWd/t
IASCC
ITER
LHGR
LMFBR
LWR
PCI
PHWR
PKA
PWR
SCC
SIA
SFT
VHTR
Self-interstitial atom
Stacking fault tetrahedron
Very high temperature reactor
- 84 -
Lent 2014
Lent 2014
Question Sheet 1
1.
2.
4A
1 A2
For isotropic elastic scattering, the average energy transferred is En 2 , and we take this
to be the energy of the primary knock-on atoms (PKAs) . The number of atoms
displaced by a PKA is given by the Kinchin and Pease analysis (see DoITPoMS TLP).
The displacement energy for iron Ed 40 eV . The cut-off energy Ec is too high to be
relevant for this problem.]
3.
Metallic uranium shows a number of problems when used as a nuclear fuel. Identify and
briefly describe those problems attributable to -U being orthorhombic (i.e. with single
crystals much less symmetric than cubic).
4.
An -U single crystal is subjected to a neutron fluence that induces the crystal to expand
by 10% along [010] (i.e. perpendicular to its (010) planes). Assuming that all of this
expansion is attributable to the development of interstitial dislocation loops on planes
parallel to (010), and ignoring all other dislocations, what is the dislocation density (in
m 2 ) induced by the irradiation?
[The spacing of the planes can be taken to be b 2 . The diameter of the dislocation loops
is 400 nm.]
- 85 -
5.
Lent 2014
Uranium dioxide is the most common nuclear fuel. Explain its relative advantages and
disadvantages compared to other possible fuels. Why does the composition of UO2
change during use? How can the effects of this change be counteracted? What potential
advantages might future fuels have?
- 86 -
Lent 2014
Question Sheet 2
1.
What are the main characteristics required in the cladding for a nuclear fuel? Which are
the main factors limiting attainable fuel burn-up?
2.
Austenitic stainless steels are commonly used for the cladding in FBRs. Under irradiation
these alloys develop voids. Explain how these voids form and why the matrix around a
void tends to become rich in nickel. What are the likely effects on the crystal structure of
the changes in matrix composition?
Mostly alloys become harder under irradiation. In which circumstances can they become
softer as a result of irradiation?
3.
Zirconium alloys are commonly used for the fuel cladding in PWRs and BWRs. Why are
they particularly suitable for this application? In use, which problems are attributable to
the particular crystal structure(s) of zirconium?
4.
Explain why the nuclides in moderators should preferably have a high cross-section for
neutron scattering and a low mass number.
5.
- 87 -
Lent 2014
Examples Class
1.
If a particle of mass A1 and energy E is incident on a stationary particle of mass A2, the
maximum energy transfer to the stationary particle is Emax = E, where:
Note that for incident neutrons, this expression reduces to that used in Q2 on the question
sheet. Assuming that the displacement energy of the target nuclide is 30 eV (a reasonable
average value), what is the necessary energy of incident (i) neutrons, (ii) protons, (iii)
alpha particles, (iv) fission fragments (average mass half that of
of mass 10, 50, 100 and 200 u? Comment on the numbers obtained.
) to displace atoms
2.
Why are austenitic stainless steels particularly susceptible to swelling under neutron
irradiation? What are the consequences of swelling?
3.
Zircaloys suffer irradiation growth, in which interstitial loops form on planes parallel to
the z-axis of the hexagonal crystal structure, and vacancy loops form on planes
perpendicular to the z-axis. Considering the coefficients of thermal expansion in single
crystals of zirconium, how can these preferred orientations of loops be explained?
In a single crystal of zirconium undergoing irradiation growth, how must the strains
parallel and perpendicular to the z-axis be related?
When voids are found in zirconium alloys under irradiation, why are they mostly near
grain boundaries?
4.
Explain the function of a moderator in a thermal nuclear reactor. What are the desirable
characteristics for the nuclides in a moderator, in terms of neutron scattering crosssection s, neutron absorption cross-section a, and mass number A? To compare the
performance of graphite, light water and heavy water as moderators, we can consider the
cross-sections of C, H, and D (see figure in the handout). In each case, for 1 MeV
neutrons arising from fission reactions, how many collisions must they undergo, on
average, before they are effectively thermalised? In the course of such thermalisation,
what fraction of the neutrons is lost by absorption? Comment on the relative suitabilities
of graphite, light water and heavy water as moderators.
5.
A control rod in a nuclear reactor must absorb neutrons; by so doing it controls the rate of
fission by inhibiting (and ultimately quenching) the chain reaction that generates
neutrons. Using the figure in the handout, and ignoring those elements likely to be used
as nuclear fuels, which elements are likely to be used in control rods?
- 88 -