3.04

Thorium Oxide Fuel

P. R. Hania and F. C. Klaassen
Nuclear Research and Consultancy Group, Petten, The Netherlands

ß 2012 Elsevier Ltd. All rights reserved.

3.04.1
3.04.2
3.04.2.1
3.04.2.2
3.04.2.3
3.04.3
3.04.3.1
3.04.3.2
3.04.3.3
3.04.3.4
3.04.3.5
3.04.4
3.04.4.1
3.04.4.1.1
3.04.4.1.2
3.04.4.2
3.04.4.3
3.04.5
3.04.5.1
3.04.5.2
3.04.5.2.1
3.04.5.2.2

3.04.5.2.3
3.04.5.2.4
3.04.5.2.5
3.04.5.2.6
3.04.5.2.7
3.04.6
3.04.6.1
3.04.6.2
3.04.6.3
3.04.7
References

Introduction
Incentives for Using Thorium
Thorium as an Abundantly Available Resource for Nuclear Fuel
Radiotoxicity Reduction with Thorium
Reduction of Excess Military Plutonium
Physical Properties of Thorium Oxide Fuel
Crystal Structure
Thermal Expansion
Thermal Conductivity
Thermophysical Properties
Oxygen Potential
Thorium Oxide Fuel Fabrication
Powder Compaction
ThO2
Mixed oxides
Powder–Plasticizer Methods
Sol–Gel Methods
Behavior of Thorium Oxide Fuel Under Irradiation

Neutronic Properties of Thorium-Based Fuel
In-Core Behavior of Thorium Oxide Fuel
Restructuring
Thermal conductivity
CANDU fuel
United States
India
Fission product behavior
(Th,Pu)O2
Reprocessing and Refabrication
The THOREX Process
Beyond Thorex
Radiation Issues in Reprocessing and Refabrication
Conclusions

Abbreviations
ANS
BWR
CANDU
fcc
FGR
HEU
HMTA
IAEA
HTR

American Nuclear Society
Boiling Water Reactor
CANadian Deuterium Uranium reactor
Face-centered cubic

Fission Gas Release
High Enriched Uranium (>20% U-235)
Hexamethylene tetramine
International Atomic Energy Agency
High Temperature Reactor

LEU
LWR
MOX
MSR
Mwe
PUREX
PWR
tHM
TBP
TRISO

88
89
89
90
90
90
91
91
92
93
93
94
94

94
95
96
96
97
97
99
100
100
101
101
102
102
103
104
104
105
106
106
106

Low Enriched Uranium (<20% U-235)
Light Water Reactor
mixed oxide
Molten Salt Reactor
MegaWatt Electric
Plutonium Uranium EXtraction
Pressurized Water Reactor
ton Heavy Metal
TriButyl Phosphate

Tristructural Isotropic

87


88

Thorium Oxide Fuel

THOREX
WG
XRD
EBWR
ETR
MTR

THORium EXtraction
Weapons Grade
X Ray Diffraction
Experimental Boiling Water Reactor
Engineering Test Reactor
Materials testing reactor

3.04.1 Introduction
Thorium (Th) is an actinide element with atomic
number 90. It is a silvery-white colored metal, discovered in 1828 by the Swedish chemist Jo¨ns Jacob
Berzelius. The element was named after Thor, the
Germanic god of thunder. Natural thorium contains
one isotope, 232Th, an a-emitter with a half-life of
1.4 Â 1010 years. In addition, other Th-isotopes are

found in nature in trace amounts, as daughter-products in the decay chains of uranium (227Th, 230Th)
and thorium itself (228Th). These isotopes are much
shorter lived than their mother isotopes and, consequently, much less abundant. As 232Th is basically the
only natural isotope, in this chapter 232Th is meant
when referring to thorium.
Although 232Th is a nonfissile isotope, it can be
used as fuel in nuclear reactors. By capturing a neutron in 232Th, 233U is formed, according to the
following nuclear reaction:
232

Th þ n !

233

Th !

Pa !

233

233

U

Protactinium ( Pa), a b-emitter with a half-life of
27.0 days, is formed as an intermediate product; the
end product, 233U, is a fissile uranium isotope. For this
reason, thorium can be used as a fertile isotope to
generate new fissile material, similar to the transmutation of 238U into 239Pu. As thorium is more abundant
than uranium, with otherwise similar chemical and

(neutron) physical properties, it opens up the possibility to include a new fissile nuclide and potentially a
way to a more efficient use of resources. In order to
start-up a thorium fuel cycle, that is, a nuclear fuel
cycle based on the fissioning of 233U bred from 232Th,
sufficient fissile material must be generated first. Initially, the breeding of 233U must be assisted by a sufficient amount of fissile material in the core of the
nuclear reactor. This can be done by either enriched
uranium (with enrichments higher than in conventional fuel) or plutonium.
The potential of thorium was identified in the early
stages of nuclear technology development. The Shippingport atomic power reactor in the Unites States was
233

fueled with thorium to establish a 232Th/233U fuel
cycle, from 1977 until its decommissioning in 1982.1
In Germany, the THTR-300, a high-temperature
reactor (HTR), operated from 1983 to 1989 with a
core consisting of HTR pebbles with thorium and
(highly enriched) uranium.
However, because uranium proved to be abundantly available to cover the world demand for the
production of nuclear fuel, the necessity to develop a
thorium fuel cycle on an industrial scale never arose.2
Nevertheless, the potential of thorium as a resource
for nuclear power has always been recognized and
researched, for example, in countries with large thorium reserves and less access to uranium resources.
India is probably the best example of a country that
has developed a comprehensive approach to a sustainable nuclear fuel cycle based on thorium.3
In the last decade of the twentieth century, a
second incentive came up, that of using thorium as
a way to reduce the radiotoxicity of spent fuel. This
radiotoxicity is dominated by the transuranium elements plutonium, americium, and curium. The use of
thorium instead of 238U as the main fertile isotope

reduces the amount of transuranium isotopes in the
spent fuel by two orders of magnitude. As a result,
the lifetime of nuclear spent fuel, i.e. the time needed
to reduce radiation levels to that of unirradiated UO2,
can be reduced significantly. A third option considered was to use thorium as a means to efficiently
incinerate excess military (weapons-grade, WG) plutonium from the disposition programs in Russia and
the United States. In fact, thorium is a suitable matrix
for plutonium.
This chapter focuses on thorium oxide as nuclear
fuel. The oxide is the most investigated chemical
form and the one most likely to be used in the existing reactor fleet (but to give other examples: ThZr
metal fuel has been developed in the United States,
and thorium mixed carbides and oxicarbides have
been studied in the United States and Germany;
see Chapter 3.02, Nitride Fuel and Chapter 3.03,
Carbide Fuel).
Much of the relevant research from the last century is available in numerous overview articles,2,4,5
and the IAEA ‘Technical Documents’6–10; these
works give a fairly complete picture but tend to
focus on reactor physics and then mostly on (Th,U)
O2 fuels containing high enriched uranium (HEU).
Two other useful sources are a comprehensive
Batelle report from 1979 assessing the available data
on properties, fabrication, and irradiation performance of ThO2–UO2 pellet fuels11 and a review


Thorium Oxide Fuel

article by Rodriguez and Sundaram from 1981 containing many references to the older literature.12
Here we give an introduction with more emphasis

on the materials perspective, including the more
recent trend of replacing the HEU with LWR- or
WG-plutonium (WG-Pu), stemming from a virtual
ban on HEU. Specifically, we treat the basic properties, fabrication and reprocessing aspects, and the
irradiation characteristics of thorium oxide fuels.
Three recent complementary articles from the
Encyclopedia of Materials should be mentioned, which
give up-to-date overviews of general fuel fabrication
issues, thorium fuel cycles, and reprocessing issues,
respectively.13–15
The chapter is organized as follows: in Section
3.04.2, the rationale behind the investigation of thorium as a fissile source material is highlighted. In
Section 3.04.3, the basic physical properties of thorium oxide fuels are discussed in comparison with the
oxides of uranium and plutonium. The fabrication
aspects are discussed in Section 3.04.4. In Section
3.04.5, the behavior of thorium oxide fuel under
irradiation is treated. Reprocessing issues are discussed in Section 3.04.6.
The authors do not claim to have presented a
complete picture of thorium in every detail. However, they do hope to have presented a thorough
overview as well as numerous references that may
help the reader further.

3.04.2 Incentives for Using Thorium
As indicated in the introduction, there are three main
incentives for the use of thorium. These are the use of
thorium as a resource alternative for uranium, the
potential of the thorium fuel cycle to reduce the
radiotoxicity of spent fuel, and a third, more specific
application of thorium as a matrix for the incineration of excess WG-Pu. These three applications are
discussed in the following sections.

3.04.2.1 Thorium as an Abundantly
Available Resource for Nuclear Fuel
Thorium is more abundant than uranium: The content of thorium in the earth’s crust is estimated to be
about three to four times larger than that of uranium.
Its most common source is the rare earth mineral
monazite. Monazite sand is found in large amounts
in India and Brazil, and Australia has large deposits
as well. It should be noted, however, that the exact

Table 1
Total identified world thorium
(in 106 kg Th), sorted by amounts per country

89

reserves

Country

Identified Th resources
(<80 USD per kg Th)

Australia
United States
Turkey
India
Brazil
Venezuela
Norway
Egypt

Russian Federation
Greenland
Canada
South Africa
Others
Total

452
400
NA
319
302
300
132
100
75
54
44
18
33
2229

Data taken from Uranium Red Book; OECD 2009.
NA ¼ not available.

amount of thorium resources is not very well known.
Estimates of thorium resources (identified resources,
retrievable at a cost less than USD 80 per kg Th) are
given in Table 1, taken from the Uranium Red Book
2009.16 This overview estimates a total of known

resources of 2.5 million tons Th (identified resources)
and an additional prognosticated amount of 1.9
million tons.
Currently, there is little demand (and industrial
need) for thorium; it is obtained mostly as a byproduct from the production of rare earth metals
or from the production of titanium-, zirconium-, or
tin-bearing minerals from monazite deposits.16 Little
systematic exploration has therefore been conducted
specifically for thorium, which explains the relative
uncertainty about the total world reserves.
The large availability and relatively easy retrievability have been a large incentive to look at thorium
as nuclear fuel.17 Specifically, this has been the case in
countries with large thorium reserves, such as India,
Brazil, and, recently, Norway.
The largest advantages of 232Th over the comparable fertile isotope 238U are the amount of neutrons produced per fission and the fission-to-capture
ratio, which are both higher than in 238U (Section
3.04.5.1). This allows breeding with thorium in thermal reactors as well, whereas for 238U, a fast spectrum
is mandatory for breeding. Thorium can thus be used
directly in light or heavy water reactors or in HTR. It
should be noted, though, that achieving breeding
with thorium is not straightforward. The exact gain


90

Thorium Oxide Fuel

in resource efficiency, which can be achieved in a
thorium cycle, as compared with the ‘standard’ uranium cycle, depends highly on the chosen fuel cycle
scenario, that is, the combination of chosen reactor

type, fuel type, and fuel management scenario. The
efficiency of breeding 233U together with the burnup
rate of the fissile topping, and the choice of whether
or not to reprocess the spent fuel, ultimately determines the efficiency of a chosen thorium fuel cycle.
The Shippingport reactor constitutes a realization
of breeding in a light water-moderated reactor, but
this was a small (60 MWe) reactor with an optimized
breeding fuel cycle. In heavy water reactors, neutron
absorption is smaller, which translates to a higher
breeding potential. The graphite-moderated HTR
offers a similar but smaller advantage over light water.
In the longer term, the use of thorium is foreseen in
molten salt reactors (MSR) as ThF4 (see Chapter 3.13,
Molten Salt Reactor Fuel and Coolant).
3.04.2.2 Radiotoxicity Reduction
with Thorium
A second incentive to look at thorium is connected
to the potential of waste production. Thorium fuels
produce far less transuranic elements (i.e., elements
with atom number >92). As these transuranic elements determine the long-term radiotoxicity of the
spent fuel, the use of thorium greatly reduces this
radiotoxicity. A common definition of the lifetime of
spent nuclear fuel is connected to the decay time, in
which the spent fuel reaches the radiotoxicity level of
the original uranium ore from which it was produced.
According to this definition, actinides (transuranics)
in the spent UO2 fuel have a lifetime of 130 000 years,
whereas fission products have a lifetime of only
270 years.18 The lifetime is thus dominated by the
transuranic elements plutonium, americium, and

curium, with atom mass numbers of 238 and higher.
These are produced by subsequent neutron captures
in 238U. The lower mass number of 232Th ensures
that these higher isotopes are produced in much
less amounts.
The potential to reduce radiotoxicity of spent
nuclear fuel with thorium was investigated extensively in Europe under the 4th Framework program
‘Thorium as a waste management option’4 and later
in the 5th Framework program ‘Thorium cycle.’108
The effect of reduction in radiotoxicity depends
greatly on the choice of ‘topping,’ which is the fissile
start-up material. At the start-up of a thorium cycle,
that is, when no fissile 233U is available, one has,

generally, the choice of three toppings, either HEU,
low enriched uranium (LEU), or plutonium (Pu). It is
clear that the use of HEU offers the most benefits in
terms of radiotoxicity reduction. But the use of HEU
in a civil nuclear fuel cycle is not preferred due to
reasons of proliferation risks. Furthermore, the effect
of resource efficiency through the use of abundant
thorium is counterbalanced by the amount of natural
uranium material needed to enrich uranium to
enrichments of $90%. The combination of thorium
with HEU, although preferred from a radiotoxicity
point of view, is therefore not a viable option for a
sustainable fuel cycle. The more the amount of uranium (238) or plutonium present in the initial fuel,
the more the amount of transuranics produced, which
determines the lifetime of the spent fuel. Nevertheless, a significant reduction of the radiotoxicity can
be achieved. Once an equilibrium thorium cycle is

achieved, the benefits increase, as then the fuel
is based on the combination of 232Th/233U only.
3.04.2.3 Reduction of Excess Military
Plutonium
A third, more specific potential application of thorium is connected to the incineration of excess
military plutonium. More than 250 tons of WG-Pu,
containing around 93% 239Pu, has been produced in
the world for military purposes, mostly by the United
States and the Russian Federation. Part of these stockpiles has been declared as excess plutonium, and both
the United States and Russia have agreed to dispose of
34 ton WG-Pu. In the disposition program, it has been
well defined how the excess plutonium will be incinerated, that is, through its use as (U,Pu)O2 fuel in
nuclear power plants in the United States and as fuel
for fast reactors in Russia. Nevertheless, the mixture
of thorium with WG-Pu into (Th,Pu)O2 fuel may
provide a technical option to reduce the threat of
military plutonium. The application of (Th,Pu)O2
fuel may similarly be used to more efficiently reduce
stockpiles of separated civil plutonium.

3.04.3 Physical Properties of
Thorium Oxide Fuel
The following section describes the important
thermophysical properties of ThO2 in comparison
with UO2 and PuO2. The comparison with the more
common fuel oxides UO2 and PuO2 (see Chapter 2.02,
Thermodynamic and Thermophysical Properties


Thorium Oxide Fuel


of the Actinide Oxides) is made throughout to show
that the three compounds are very similar and to simultaneously highlight the differences, which turn out
mostly favorably for ThO2. What sets ThO2 fuel apart
is that the thorium ion (unlike uranium and plutonium)
adopts in compounds a single oxidation state (4þ). This
single valence implies that the oxide has very little
nonstoichiometry, that is, a low number of oxygen
vacancies and interstitials. The absence of nonstoichiometry is reflected in a relatively high thermal conductivity (Section 3.04.3.3), a well-defined oxygen
potential (Section 3.04.3.5), high thermal stability,
low chemical reactivity, low matrix diffusivities, and
some difficulty in the sintering of pellets (Section
3.04.4). For more thorough discussions of the existing
literature on thermophysical properties, we refer the
reader to two exellent reviews20,21 and a paper by
Sobolev et al. discussing some relationships between
the different properties.123
3.04.3.1

3.04.3.2

Thermal Expansion

Touloukian et al. recommend the following relation
for the thermal expansion of pure ThO234 based on
a large set of measurements in excellent agreement:
DL=L0 ð293 KÞ ¼ À0:179 þ 5:097 Â 10À4 T
þ 3:732 Â 10À7 T 2
À 7:594 Â 10À11 T 3 ð150 À 2000 KÞ
The thermal expansion coefficient aL is obtained by

differentiation. Figure 1 compares this expansion
for ThO2 with that observed for UO2, and PuO2,
which have all been described using a single set of
relations25:

A comparison of physical parameters for the dioxides of Th, U, and Pu

Property

Unit

Lattice constant, a
Cell volume, Vc
Molecular weight, Mw
Theoretical density, r
Melting temperature, Tm
Standard enthalpy of formation, DH0
Standard entropy, S0
Heat of fusion, Hf

´
A˚
˚A´ 3
a.u.
g cmÀ1
K
J molÀ1
J molÀ1 K
J molÀ1


a

changes linearly with the additive fraction (Vegard’s
Law behavior). However, locally the individual heavy
metal–oxygen bond lengths in the mixed oxide
stay somewhat closer to the values of the pure
compounds.30,31
It follows from the Vegard’s Law behavior of
the stoichiometric mixed oxide that their roomtemperature densities may simply be obtained by
linear interpolation of the weights and cell volumes
given in Table 2 and additionally that the solid solutions thus formed show ideal behavior.20 Vapor pressure measurements on ThO2–UO2 solid solutions
have indicated that the same ideality also holds at
higher temperatures.32,33 This allows one to reverse
the argument and claim that the lattice parameter
also changes linearly with composition in the hightemperature region. Therefore, the thermal expansion of the mixed actinide oxides can be obtained by
linearly interpolating the thermal expansions of the
pure compounds.20

Crystal Structure

Close similarities exist between the three actinide
oxides: In common with UO2 and PuO2, ThO2 possesses an fcc lattice, and all three materials can be
heated to melt without undergoing phase transitions.20 In addition, thermal conductivities and thermal expansions are similar. The three oxides may
furthermore be mixed in all proportions, forming a
single-phase material.
The associated lattice constant for this series of
actinides decreases as Th < U < Pu, as shown in
Table 2, which compares some physical parameters
for the stoichiometric compounds.
XRD measurements indicate that when changing

the composition from pure ThO2 to pure PuO227 or
UO2,28,29 the lattice parameter of the cubic lattices
Table 2

91

ThO2.00
(fcc)

UO2.00
(fcc)

PuO2.00
(fcc)

5.597
175.3
264.04
10.00
3651 Æ 17
À1226.4
65.23
90

5.4702
163.7
270.03
10.96
3120 Æ 30
À1085.0

77.03
78

5.3815
157.5
276
11.46
3017 Æ 24a
À1055.8
66.13
67 Æ 15

Value taken from a recent laser flash measurement,124 but note that PuO2 melts about 300 K below this point in older studies.25
The molecular weight of Pu, which has no stable isotopes, is conventionally fixed at 244. Lattice constants, cell volumes, and melting
temperatures are taken from IAEA Technical Document20 and thermophysical data from Bakker et al.19 and Cordfunke and Konings.23


Thorium Oxide Fuel

14

ThO2

2.0

ΔL/L (%)

UO2, PuO2
1.5


1.0

0.5

0.0

400

600

800 1000 1200 1400 1600 1800 2000
Temperature (K)

Figure 1 Thermal expansion of ThO2 and UO2/PuO2. In
the range shown, both relations are valid. The expansion
curve for ThO2 has been adjusted here to a reference
temperature of 273 K.

DL=L0 ð273 KÞ ¼ À0:266 þ 9:802 Â 10À4 T
À 2:705 Â 10À8 T 2
þ 4:391 Â 10À11 T 3 ð273 À 923 KÞ
DL=L0 ð273 KÞ ¼ À0:328 þ 1:179 Â 10À3 T
À 2:429 Â 10À7 T 2
þ 1:219 Â 10À10 T 3 ð923 K À Tm Þ
where Tm is the melting temperature. The comparison shows that thermal expansion is slightly lower for
ThO2. For a detailed discussion of temperaturedependent theoretical density, linear expansion, and
melting points of ThO2 and ThO2–UO2 mixtures, we
refer to a recent assessment in the IAEA Technical
Document.21
3.04.3.3


Thermal Conductivity

At moderate temperatures, where the electronic
contribution can be neglected, empirical values for
the temperature-dependent thermal conductivity of
ionic solids may be fitted with the general function
l = 1/(A þ BT). Here, the constant A describes
the effect of material defects that are present
independent of temperature and act as phonon scattering centers, while the term BT represents the
temperature-dependent effect of phonon–phonon
interactions.35 Bakker et al.20 have analyzed a large
set of experimental data and used a selection to obtain
values of the parameters A (4.20Â10À4m K WÀ1) and
B (2.25Â10À4m WÀ1) for 95% dense ThO2. More
recent experimental data are available from Pillai36
and Cozzo.37 The three results are in reasonable
agreement (Figure 2).

Thermal conductivity (W m K–1)

92

Bakker, 95% TD
Pillai, 94% TD
Cozzo, 95% TD
UO2, Lucuta-Ronchi, 95% TD

12
10


UO2, Fink, 95% TD
PuO2, Cozzo, 88% TD

8
6
4
2
0
400

600

800
1000
1200
Temperature (K)

1400

1600

Figure 2 Thermal conductivity of stoichiometric ThO2,
UO2, and PuO2.

In Figure 2, a comparison is also made with the
correlations given for UO225,38,39 and PuO2.37 The
thermal conductivity of UO2 is obviously below that
of ThO2 in the temperature region shown in this
figure (up to 1600 K); however, an electronic contribution to the thermal conductivity of UO2 kicks in at

temperatures above 2000 K, whereas this contribution is absent for ThO2.40
For PuO2, Cozzo et al. argue that previous studies41,42 yielding values close to those for UO2 had
most probably been performed on samples of illdefined stoichiometry and that stoichiometric PuO2
in fact has the largest thermal conductivity of the
three oxides.37 This argument indicates that due to
the multivalence of the metal atoms, control over stoichiometry is not trivial. Even when the stoichiometry is
carefully controlled during fabrication, PuO2 and UO2
will become nonstoichiometric upon heating or when
in oxidative or reductive environments. This results,
for many practical conditions, in some loss of thermal
conductivity, which is difficult to control.
Mixing of the actinide oxides generally depresses
the thermal conductivity, mostly because the additive
heavy metal ions act as phonon-scattering centers.
The scattering term A is therefore affected more
strongly than the phonon interaction term B. Gibby
has observed this trend for the mixing of uranium and
plutonium dioxide,41 and on ThO2 the depression is
quite pronounced.43
Using a small dataset from Murabayashi44 and
Springer and Lagedrost45 selected from the available
literature, Bakker et al.20 derive values of the parameters
A and B for 95% dense Th1ÀyUyO2 and y up to 0.1
(T ¼ 300–1800 K): A ¼ 4.195 Â 10À4 þ 1.112y À 4.499y2
(m K WÀ1),
B ¼ 2.248 Â 10À4 À9.170 Â 10À4y þ
À3 2
4.164 Â 10 y (m WÀ1). These authors reject on



Thorium Oxide Fuel

93

theoretical grounds an earlier assessment by Berman
et al.46 from a partially overlapping dataset; however,
Berman gives correlations for higher uranium concentrations. According to Bakker’s recommendation,
the addition of 10% UO2 to ThO2 results in a reduction of the conductivity by 40–50%. More recently,
Pillai et al. have found that the addition of as little as
2% UO2 decreases the thermal conductivity of thoria
by 10–30%.36 In the recent IAEA review,21 new data
are considered along with selected literature to
obtain A and B values for uranium contents of 0, 4, 6,
10, and 20%. Compared with Bakker’s analysis, a
stronger temperature dependence is obtained in this
work (the conductivities being equal at about 1400 K).
In addition, the effect of uranium concentration is
smaller, that is, for a given uranium concentration,
the conductivity depression is smaller.
For the addition of PuO2, few sources are
available, but Cozzo et al.37 report the correlation
A ¼ 6.071 Â 10À3 þ 0.572y À 0.5937y2
(m K WÀ1),
À4
À1
B ¼ 2.4 Â 10 (m W ). This correlation has a minimum at around 45% Pu, at which point the thermal
conductivity has been reduced by more than a factor
2 at 500 K. An equation obtained from a different
dataset is given in21: A ¼ À0.08388 þ 1.7378Â10À4y
(m K WÀ1), B ¼ 2.62524 þ 1.7405Â10À4y (m WÀ1).

For a plutonium concentration of 5%, this result is
about equal to that of Cozzo, but the effect of Pu
addition is found to be stronger. Basak reports similar
findings for the addition of 4% PuO2.47
The above comparison of the available correlations
for (Th,U)O2 and (Th,Pu)O2 fuels reveals significantly
different results, which seems to indicate measurement
uncertainties related to sample microstructure and
stoichiometry. When using the Berman40 and Bakker20
correlations for (Th,U)O2 and the Cozzo37 correlation
for (Th,Pu)O2, the fissile concentrations at which
the thermal conductivity of thoria-based fuel becomes
equal to that of UO2 is about 10%. However, the
correlations given in a recent assessment in an IAEA
Technical Document21 suggest that the thermal conductivity of UO2 is approached with the addition of
about 20% uranium or only 6% Pu.

The standard entropy, based on measurements
of the low-temperature heat capacity, is
S 0(298.15) ¼ 65.23 J KÀ1 molÀ1.50 Bakker et al.20 have
fitted earlier measurements of H(T)ÀH0(298 K)
under the constraint Cp0 (298 K) ¼ 61.76 J KÀ1Ámol
from the low-T heat capacity50 to obtain a function
for the integrated high-temperature heat capacity:

3.04.3.4

3.04.3.5

Thermophysical Properties


The standard enthalpy of formation DfH (298.15)
is À1226.4 kJ molÀ1 (Table 1),48 which makes ThO2
the most stable oxide known. As shown in Table 2,
this is reflected in a significantly increased melting
temperature with respect to UO2 and PuO2 (recommendations by Ronchi et al.20,38 and Martin et al.25,49).
0

H 0 ðT Þ À H 0 ð298Þ ¼ 55:962T þ 25:62895 Â 10À3 T 2
À 12:2674 Â 10À6 T 3
þ 2:30613 Â 10À9 T 4
þ 5:740310 Â 105 T À1
À 20581:7
The given temperature dependence shows an excess
enthalpy at temperatures above 2500 K, similar to that
of other actinide oxides, which has been related to a
premelting l-transition at 3090 K.38 Similar discontinuities in the slopes of the enthalpy–temperature
curves at about 0.8 of the melting temperature were
found for a series of thoria–urania solid solutions.51
Recently, Dash et al. have performed extensive
measurements on Th1ÀyUyO2 for y 0.2 (127–
1698 K), making a comparison to older data.52 They
arrive at the following expression for the heat capacity as a function of T (298 T 2000 K) and
y (0.019 y 0.9):
Cp ¼ ð66:26 þ 10:91y Þ þ ð0:00923 À 0:00065y ÞT
À
Á
À 7:70 Â 10À5 þ 6:7 Â 10À5 y T 2
The correlation of Dash et al. for pure ThO2 corresponds reasonably well to that obtained by Bakker et al.
With regard to the mixed oxides, it can be said that

both (Th,Pu)O2 and (Th,U)O2 experience a slight
melting point depression. For (Th,U)O2, the melting
points correspond to that of ideal solid solutions;
for (Th,Pu)O2, not enough data exist to confirm this
ideality.20 The Cp(T) data for the (Th,U)O2 are quite
well reproduced by a weighted average of the Cp(T)
values of ThO2 and UO2, with little deviation.20,52
The IAEATechnical Document21 uses Bakker’s correlation for ThO220 and Fink’s correlation for UO239 to
construct Neumann-Kopp heat capacities.
Oxygen Potential

As pure ThO2 is for practical purposes a stoichiometric
compound, the oxygen potential ÀRT ln(pO2) is only
a function of temperature. Using the entropy and
enthalpy values described in the previous section,
the oxygen potential at 1000  C is found to be
around À700 kJ molÀ1, significantly lower than for


94

Thorium Oxide Fuel

DG ðO2 Þ ¼ 0:029592T Á lnðx Þ þ 0:003436T
Á lnð0:5619 À 0:1161x Þ À 0:033028T

–50

DG(O2) (kJ mol–1)


–100

y = 0.05
y = 0.1

Á lnð0:5619y À 1:1161x Þ À 348
À
Á
þ 0:1767T kJ molÀ1

–150
y = 0.2
–200
–250
Schram
Dash

–300
–350
0.000

0.005

0.010
x

0.015

0.020


Figure 3 Oxygen potentials of Th1ÀyUyO2+x at 1473 K
and different values for x and y, as determined by Schram53
based on a Lindemer–Bessman model for UO2+x (black),
and as determined by Dash et al.52 from direct
measurements on the mixed oxides (red).

PuO2 and UO2 under realistic conditions. We can
combine this information with the ideality of the solid
solution and claim that thorium in mixed oxides can be
regarded as an inert solvent (for PuO2 or UO2) that
does not take part in the chemical equilibria describing
the oxygen potential.53 This reduces the oxygen chemistry of mixed thorium oxides to the chemistry of the
fissile additive (U or Pu).
On a microscopic level, hyperstoichiometry in
ThO2–UO2 solid solutions manifests as interstitial
oxygen. Cohen shows that at 1200  C, the maximum
amount of interstitial oxygen increases from 0 for
pure ThO2 to 0.25 for Th0.1U0.9O2.25.54 This increase
with U concentration follows from the fact that a
valency change from 4þ to 5þ in two U atoms or
(less likely) from 4þ to 6þ in one U atom is needed
to compensate for the presence of the extra oxygen.
The incorporation of PuO2 in the ThO2 matrix similarly allows for the creation of oxygen vacancies at
elevated temperatures under inert or reducing atmospheres. It may be clear that, because the additive
(UO2 or PuO2) will be fissioned away during irradiation and the thorium ions cannot undergo a valency
change, the oxygen potential in thorium mixed
oxides fuel should in fact rise faster with burnup
compared to conventional (U,Pu)O2.
Schram53 has used the above reasoning to describe
(Th,U)O2+x with a Lindemer–Bessman type model

for the uranium, which describes the collected (limited) amount of oxygen potential data reasonably
well. On the other hand, Dash et al.52 have used
their heat capacity measurements on Th1ÀyUyO2 to
calculate the oxygen potential in a more direct way52
(x 0.024, y 0.2):

The correlations derived by Schram and Dash et al.
are shown in Figure 3; the data are in reasonable
agreement, considering that they are based on oxygen
potential measurements55–57 and heat capacity measurements,52 respectively.
Finally, we note that phase separation occurs for
off-stoichiometric conditions at high U and Pu concentrations (Section 3.04.4.1). For instance, for high
U concentrations phase separation occurs, and the
mixed fcc oxide exists in equilibrium with a separate
U3O8 phase.54 Such a phase equilibrium would introduce an oxygen potential plateau.

3.04.4 Thorium Oxide Fuel
Fabrication
Fabrication of thorium-based oxide fuel is well developed. Three routes have been applied successfully to
create thorium-based oxide fuel: conventional binderless powder pressing, spheroidization of powder–
plasticizer mixtures, and the sol–gel process. The
latter two cases yield microspheres with diameters
in the range 50–1000 mm, which can be pressed into
pellets, used directly in Sphere-Pac/Vipac arrangements (see Chapter 3.11, Sphere-Pac and VIPAC
Fuel), or coated with carbon and silicon carbide
layers to create TRISO fuel for HTR (see Chapter
3.06, TRISO Fuel Production). Much of
the available information is from the Indian experience in powder pressing58 and from the German–
American developments in sol–gel methods.2,4,6–10
3.04.4.1


Powder Compaction

3.04.4.1.1 ThO2

The procedures for fabricating ThO2 pellets by powder compaction are derived from the procedures
developed for UO2 and (U,Pu)O2 fuel.2,10 However,
as thorium is found only as a 4-valent cation, it is not
as important to control the oxygen potential during
sintering as in the case of uranium or plutonium, and
sintering of thorium oxide may be performed in both
oxidizing and reducing conditions (air, argon, vacuum, or Ar/H2). On the other hand, the thermal and
chemical stability of ThO2 discussed in the previous


Thorium Oxide Fuel

section somewhat decreases its sinterability, and high
densities are more difficult to produce.
The thorium dioxide powder is usually obtained
by calcination of the oxalate, Th(C2O4)2, which precipitates from a nitrate feed solution (nitric acid with a
pH $0.8) upon dropwise addition of oxalic acid:
ThðNO3 Þ4 ðaqÞ þ 2C2 O4 H2 ðaqÞ
! ThðC2 O4 Þ2 ðsÞ þ 4HNO3 ðaqÞ
Early work on powders produced through different
routes (direct ignition of the thorium nitrate, decomposition of the hydroxide or carbonate) resulted in
large-grained powders that did not sinter well.
Calcination is performed in air at 800–900  C.
The ThO2 grains produced upon calcination are
fine (typically around 1 mm) and have a platelet

geometry that makes it hard to obtain high-density
sintered pellets.6,7,10 Premilling improves the sinterability considerably, but White et al. have observed
that pellets with a density of 96% TD can also be
prepared without premilling when the oxalate precipitation step is carried out below room temperature
(typically 0–10  C).59
With regard to the compaction step, it was found
that both green and sintered densities increase with
pressure for compacting pressures in the range 40–
280 MPa. However, the variations in sintered densities
(1600  C, Ar/H2) for batches of pellets were smallest
when applying pressures in the range 90–120 MPa.6
In India, a precompaction stage was introduced to
avoid chipping or breaking of the green pellets and
to increase the density of the sintered pellets. Following precompaction at around 100 MPa and subsequent granulation, the obtained granules were sieved
through a À14 mesh.6,10 Final compaction of ThO2
pellets could then be performed at higher pressures
(200–300 MPa).
Several additives have been found to considerably
improve the sinterability of the pellets.10 The idea
behind addition of sintering aids is substitution of
some of the Th4+ ions by metal ions having a different valency. The substitution introduces oxygen
vacancies or interstitials, which enhances the diffusion of thorium ions thereby producing more homogeneous and higher-density pellets.7,58 Ca2+ or Mg2+
(added at $1 wt% to the feed solution as a sulfate or
nitrate yielding $0.5 wt% in the oxide) or 0.25 wt%
Nb5+ (as Nb2O5) have thus been found to significantly
reduce the required sintering temperatures, from
1600 to 1700  C60 down to 1150–1450  C.10 As the
divalent additives introduce oxygen vacancies, they

95


tend to be better sintering aids under a reducing
atmosphere,58 while the niobate functions best in an
oxidizing atmosphere. We refer the reader to Kutty
et al.58 for an analysis of the effects of dopants.
Finally, it is well known that water easily adsorbs
to the ThO2 surface, chemically by forming a high
concentration of hydroxyl groups and physically via
hydrogen bonds.61–63 Care should therefore be taken
to store the resulting pellets in dry conditions.
3.04.4.1.2 Mixed oxides

In LWRs, around 3–5% of the heavy metal nuclides
in fresh fuel are fissionable. This means that to
replace standard UO2 or (U,Pu)O2 fuel in LWRs,
roughly 25% of LEU or 8–9% LWR-grade Pu or
5% WG-Pu or HEU should be added as a ‘topping’ to
the thoria matrix. The mixed oxide may be prepared
simply by mixing the separate oxide powders. To
enhance homogeneity of the pellets, the thorium
nitrate solution can be mixed with either uranyl
nitrate or plutonium nitrate before coprecipitation
by the addition of oxalic acid or bubbling of NH3.
Before carrying out the coprecipitation step, uranylnitrate should first be reduced by the addition of
hydrazine64 or by hydrogen gas in the presence of a
Pt/Al2O3 catalyst. For (Th,U)O2 with significant
amounts of UO2, the optimum calcination temperature with respect to sintering behavior shifts from 800
to 900  C to around 700  C.65
The fact that uranium and plutonium are multivalent suggests that the (Th,U)O2 and (Th,Pu)O2
mixed oxide are more easily sintered than pure

ThO2 under oxidative and reductive atmospheres,
respectively. Indeed, the addition of 2 wt% U3O8 to
ThO2 was found to lower the sintering temperature
in air to 1100  C, as was observed for Nb2O5; PuO2
yields results similar to Ca2+ in reducing
environments.
In Section 3.04.3.1 it has already been mentioned
that the stoichiometric (Th,U)O2 and (Th,Pu)O2
forms a single fcc phase in the entire composition
range, but that in practice off-stoichiometry resulting
from fabrication conditions results in phase separation at high Pu or U content. This should be avoided
as, for instance, the U3O8 phase has a 30% higher
volume than the fluorite phase, which results in grain
boundary separation and powdering of the fuel.66
Kutty et al. have studied the sintering behavior of
the (Th,U)O2.58,64 For ThO2–PuO2 with Pu contents
up to 30% sintered in Ar or Ar/H2, the sintered
pellets were found to be monophasic, whereas for
Pu contents of 50% and 75%, the ThO2–Pu2O3 bcc


96

Thorium Oxide Fuel

phase was found to be present besides the ThO2–
PuO2 fcc structure.58 Similarly, XRD data for both
ThO2–30%UO2 and ThO3–50%UO2 sintered in air
revealed the presence of small amounts of U3O8
(nearly hexagonal).10,64

3.04.4.2

Powder–Plasticizer Methods

The pellet fabrication route as described above is
usually performed without the addition of binder materials. India is developing a method to produce microspheres for HTR TRISO particles (see Chapter 3.06,
TRISO Fuel Production) or sphere-pac (see Chapter
3.11, Sphere-Pac and VIPAC Fuel) fuel from ThO2
powder in which a binder material is added. In this socalled CAP (coated agglomerate pelletization) process,
ThO2 powder is mixed with a plasticizer (e.g., a paraffinum/petrolatum mixture) at slightly elevated temperatures, after which the plastic mixture is simply
molded in an extruder and subsequently in a spheroidizer to form small spheres.67–69 In a second step, the
ThO2 spheres may be coated with a layer of the fissile
material. This method should minimize dust formation
as well as the number of steps to be performed under
shielded conditions (Section 3.04.6.3).
3.04.4.3

Sol–Gel Methods

The sol–gel process offers an alternative to the conventional powder mixing technology, which may be
automated and is dust-free, thereby offering a strong
advantage in radiation safety, which is especially
important when handling irradiated thorium (Section
3.04.6.3). The main disadvantage is that detailed control of the process is rather complex, which may give
problems when scaling up the process.
As is the case with the powder compaction process, this technique starts from nitrate feed solutions
of heavy metals (Th(NO3)4, Pu(NO3)4, and/or
UO2(NO3)2), although the used concentrations are
somewhat higher (2–3 M). But instead of adding
oxalic acid to induce precipitation, droplets of the

chosen heavy metal solution are exposed to ammonia, which induces the formation of microcrystallites
and thereby gelation of the sol. The resulting gel is
washed and dried, producing microspheres. After
calcination and sintering, the sol–gel microspheres
(with diameters in the range 50–1200 mm) may be
crushed and pressed into pellets or alternatively
used as is in a sphere-pac column. In the case of
HTR fuel, coatings are applied to the microspheres
to produce the well-known TRISO particles, as
is done for UO2 fuel. As in the powder–pellet

route, 1 wt% of Ca(NO3)2 may be added to the
heavy metal solution as a sintering aid, while 30 g
carbon black per mol heavy metals may be added
to produce spherical pores during the calcination
step.6–8
Several sol–gel routes have been applied in
the past, of which two have been most successful.
The KEMA internal gelation process developed
in the Netherlands for uranium and plutonium has
been adapted for thorium in India and Germany,70,71
while the external gelation of thorium (EGT) or KFA
process was developed in Germany (Ju¨lich)6 and the
United States (Oak Ridge).
In the internal gelation process, the nitrate feed
solution is mixed with a solution of hexamethylene
tetramine (HMTA, (CH2)6 N4) and urea (CO(NH2)2)
of similar concentration at a temperature of around
0  C 71–73; upon mixing at this low temperature, the
heavy metal ions form complexes with urea. The

resulting mixture is dispersed as fine droplets by a
hollow vibrating needle (frequency in the order
102–103 Hz). The dispersed droplets fall into a hot
(50–90  C) bath of silicone oil, and the droplet temperature rises quickly. This causes decomposition of the
heavy metal–urea complexes as well as hydrolytic
decomposition of HMTA. The latter produces ammonium hydroxide. After hydroxylation of the heavy
metal ions by NH4OH, the resulting heavy metal
hydroxides form agglomerates of microcrystallites.
This induces the sol to gel transition.
Details of the reaction for thorium (and also for
(Th,U)O2 with up to 10% U) have been described by
Kumar71 and Pai et al.74 In brief, the concentrated Th
(NO3)4 solution is partially preneutralized by adding
formaldehyde or ammonia. The optimum Th4+ concentration after mixing is 1–1.4 M, and the ratio of both
HMTA and urea to Th4+ ions is 1.4:1 (Figure 4). The
formation of opaque hard gel spheres upon gelation at
a temperature of 60–70  C is taken as an indication that
crack-free spheres will be formed after drying and
calcination.71–73 Following the gelation step, the gels
are prewashed using CCl4 to remove the oil.
In the external gelation method, no ammoniaproducing additive (HMTA) is used to hydrolyze or
polymerize the heavy metal sol.70 Instead, after being
released from the vibrating needle, the sol drops pass
through gaseous ammonia, which quickly gels the surface of the droplet. The partially gelled drops fall into a
solution with a pH of $8 containing 1% ammonia and
4 M NH4NO3, which completes gelation of the inside
of the droplets. As for the internal gelation process, the
best results were obtained when the heavy metal



Thorium Oxide Fuel

3.04.5.1 Neutronic Properties of
Thorium-Based Fuel

1.9

1.8

1.7

Region B:
opaque
hard gel
(60–70 ЊC)

Region C:
transparent
hard gel
(50–60 ЊC)

HMTA and urea (M)

1.6

1.5
Region A:
opaque soft gel
(60–75 ЊC)


1.4

1.3

1.2

1.1

1.0
0.9

97

Cracked during processing
Yielded defect-free microspheres
Too soft for processing

1.0

1.1

1.2
1.3
Thorium (M)

1.4

1.5

Figure 4 Gelation field diagram for ThO2 as published

by Kumar.68

solution (90  C) was ‘preneutralized’ up to a pH,123 of
3.25–3.5 and a viscosity of 0.03 Pa s before gelation.70,122
Treatment of the gel spheres is very similar in the
two routes described. The spheres are ‘aged’ and
washed in a 1% NH3 solution, to improve the internal
structure and to remove organic material, respectively.
The particles are then dried at 100–400  C in humid
air to avoid crack formation, calcined at around 700  C
in air, and finally sintered at 1000–1200  C. The sol–gel
particles sinter very well due to the small crystallite
size of the oxide formed from the agglomerate of
hydroxide microcrystallites. The particles thus reach
a density close to the theoretical density.

3.04.5 Behavior of Thorium Oxide
Fuel Under Irradiation
For in-depth treatment of the effects of irradiation on
UO2 and MOX fuel, see Chapter 2.17, Thermal
Properties of Irradiated UO2 and MOX; Chapter
2.18, Radiation Effects in UO2 and Chapter 2.19,
Fuel Performance of Light Water Reactors (Uranium Oxide and MOX).

Although uranium is ‘directly’ fissile, one should bear
in mind that the fissile content (235U) in natural
uranium is only 0.7%. In most reactor types, enrichment is needed before it can be used to generate
energy. 232Th should instead be compared directly
with 238U. Both are fertile isotopes that are converted
to fissile material in the reactor core. In the case of

232
Th, the resulting fissile isotope is 233U; in 238U, it
239
is Pu. It is therefore useful to compare the neutronic properties of the two sets of nuclides and
discuss their differences. The top and bottom panels
in Figure 5 compare the neutron capture cross-sections
of fertile nuclides and the fission cross-sections of
fissile nuclides, respectively. In Table 3, some neutronic properties of the relevant nuclides are listed;
the table is an abstract from Kaye and Laby.75
Regarding the fertile materials, thermal capture is
almost 3 times higher for 232Th than for 238U, but
resonance capture is more than 3 times higher for
238
U. In the fast region of the spectrum, the crosssections are similar. More important for thermal
breeders are the characteristics of the fissile nuclides.
It can been seen in Figure 5 that the fission crosssection of 233U is least dependent on neutron energy,
being relatively small in the thermal region and
relatively large in the epithermal and fast regions.
Table 3 shows that the resonance integral for
233
U fission is more than 2 times larger than that for
239
Pu. However, the most significant advantage of
233
U compared with 235U and 239Pu is the very high
fission-to-capture ratio (sf/sc), which in a thermal
spectrum is about 10 for 233U but only about 2.5 for
239
Pu.2 This produces a high neutron yield per
absorption  ¼ nsf/(sf þ sc) in a thermal spectrum

and up to energies of about 100 keV, that is, it yields a
relatively good neutron economy in a wide range of
spectra and especially in thermal and epithermal
reactors. The value of  also decreases less with
temperature compared to 235U and 239Pu.
On the other hand, a significant drawback of the
Th–U cycle is the larger time constant for b-decay of
the intermediate species: 27 days for 233Pa compared
with 2.3 days for 239Np. This has consequences for
reactor physics and handling of spent fuel. Th-fueled
reactors with high neutron densities would contain
significant levels of 233Pa due to its slow decay, and
because it possesses a relatively high absorption
cross-section, the protactinium will act as a ‘neutron
poison.’ The thorium cycle therefore benefits from


98

Thorium Oxide Fuel

10 000
1000

s (barn)

100
10
1
0.1

0.01
1E - 3
1E - 4
1E - 12

232Th
238U

1E - 10

1E - 6

1E - 8

1E - 4

0.01

1

100

0.01

1

100

E (MeV)
1 00 000

10 000

s (barn)

1000
100
10
233U

1

0.1
1E - 12

235U
239Pu

1E - 10

1E - 6

1E - 8

1E - 4

E (MeV)
Figure 5 Capture (top) and fission (bottom) cross-sections of relevant nuclides.
Table 3

Nuclear data of relevant actinides

Average over thermal spectrum

Resonance integrals

sc

sc

Fissile materials
233
U
42.20
235
U
86.70
239
Pu
274.32
241
Pu
334.11
Fertile materials
232
Th
6.533
240
Pu
262.65
238
U

2.414

sf

n

468.2
504.81
699.34
936.65

2.495
2.433
2.882
2.946

134.16
131.97
184.06
169.13

–
6.13 Â 10À2
1.05 Â 10À5

–
2.784
2.489

84.97

8448.7
277.7

Average over fission
spectrum

sf

n

sc

sf

n

751.71
271.53
289.36
570.66

2.498
2.438
2.876
2.933

0.063
0.095
0.065
0.226


1.841
1.219
1.8
1.626

2.596
2.583
3.091
3.151

–
3.74
2.16 Â 10À3

–
2.785
2.49

0.102
0.095
0.07

0.071
1.349
0.3

2.093
3.013
2.598


Source: Data taken from Tables of Physical & Chemical Constants. 4.7.2 Neutron cross-sections. Kaye & Laby Online, version 1.0, 2005;
www.kayelaby.npl.co.uk.
c is the capture cross-section, f the fission cross-section, and  is the average amount of neutrons produced per fission.
Cross-sections are given in barns (10À24 cm2).


Thorium Oxide Fuel

systems with low neutron densities and from systems
with on-line removal of fission products.
In a fast spectrum, the advantages of 233U over
235
U and 239Pu are in large part lost or countered
by some disadvantages: First, 239Pu produces more
neutrons than 233U in the fast part of the spectrum,
the latter value staying relatively constant over the
spectrum. Second, the ‘fertile’ material 238U is much
more fissionable than 232Th in this region, upping the
neutron yield. And finally, the capture cross-section
advantage of 232Th over 238U is strongly reduced
(see Figure 5).2
One of the most important characteristics of thorium nuclear fuel from a radiological point of view is
the formation of 232U. 232U is generated in thorium fuel
via several routes, but mostly from 233U through
(n,2n)-reactions. It has a relatively short half-life of
69 years and a multistep decay chain ending at the
stable 208Pb. One of the daughters, 208Tl, is radioactive,
with a hard g of 2.6 MeV. Therefore, the presence of
232

U, even in small amounts, in spent Th-fuel complicates handling during fabrication and reprocessing
considerably. It is therefore an important consideration
in the fuel cycle approach to minimize the amount of
232
U in spent fuel or alternatively to opt for a oncethrough application of thorium, without reprocessing.
At the same time, the intense g-rays produced by the
spent fuel enhance proliferation resistance.
We conclude that from the perspective of neutron physics, the Th–U cycle generally compares
well with the existing U–Pu cycle and that a significant breeding advantage exists for thermal reactors.
The current prevalence of uranium-based fuels in
thermal reactors is therefore not due to neutronic
properties of the fissile and fertile components but
mostly for historic reasons and perhaps due to the
absence of a fissile component in naturally occurring thorium.
3.04.5.2 In-Core Behavior of Thorium
Oxide Fuel
Much experience with thorium oxide fuel has been
gained in the past. Several reactors have run at least
partially on ThO2-based fuel: the AVR and the prototype THTR in Germany, (both graphite-moderated
HTRs), the Elk River, Indian Point 1, Peach Bottom
HTGR and Shippingport LWBR reactors in the
United States, and, most recently, the heavy watercooled power stations Kakrapar-1and Kakrapar-2 in
India. In addition, several irradiation tests have
been performed, a.o. in the Dragon HTR in England,

99

the Lingen BWR in Germany, the CANDU-prototype
NPD and the NRX and NRU research reactors in
Canada, and the CIRUS reactor in India.76

In most of these irradiations and especially in the
earlier ones, the fissile component in the fuel has
been HEU. Some more recent work has focused on
plutonium-containing fuels: two relatively recent
projects on this topic (OMICO and Thorium Cycle)
have been performed within the 5th Framework program of the European Commission,77 while several
thorium–plutonium irradiations have been carried
out in India and Canada.8
It is instructive to make the following points based
on data for the unirradiated materials (Section
3.04.3), before discussing the available data on actual
in-core behavior of thorium mixed oxide.
 The higher melting point and somewhat higher
thermal conductivity of ThO2 fuels compared
with those of UO2-based fuels result in a larger
temperature margin to fuel melting. This is accompanied by enhanced physical stability, which adds
to safety under accident conditions.
 The larger margin ‘to melt’ is also an indication
that all thermally activated processes (creep, oxygen diffusion, and migration of fission gases and
volatile fission products) proceed more slowly than
in UO2-based fuel.78 The very low concentration
of oxygen defects in ThO2, and the slightly
reduced temperature gradient in the pellets add
to this effect. This slowdown of thermal processes
has a few notable consequences:
– a reduced chance of cladding corrosion by corrosive fission products79
– lower fission gas release compared with UO2based fuel. This reduces rod internal pressure
(especially important under accident conditions) as well as the loss of gap conductance
due to the heavy fission gases Xe and Kr.80 On
the downside, higher retention of fission gases

could results in increased fuel swelling and
stronger fuel–cladding interaction.81
 The high chemical stability of ThO2 suggests that
the fuel is more resistant to corrosion in case of
coolant ingress.
 The net result of uranium fissioning and oxidation
of some of the fission products is a gradual increase
of the effective O/M ratio in the fuel. The inability
of ThO2 to be further oxidized, together with the
slow removal of oxygen dissolved in the matrix,
could result in a significantly higher oxygen potential in the fuel.81 As a result, it is likely that more


100

Thorium Oxide Fuel

fission products are present in the fuel in oxidized
states. For example, molybdenum, which has an
oxidation potential slightly above that of fresh
fuel, will start to oxidize even during the initial
phases of the irradiation,81 whereas in the case of
UO2, this occurs only at higher burnups.
The advantages of ThO2 over UO2 inferred above
are expected to diminish with increasing uranium or
plutonium content at varying rates. The fact that
many quantitative uncertainties remain may be illustrated by the case of postirradiation thermal conductivity. In Section 3.04.3.3, a remark has been
made on the uncertainties in measurements on fresh
fuel. Irradiation effects introduce additional uncertainty through the following effects:
 A decrease in thermal conductivity during the early

stages of irradiation due to the formation of lattice defects by neutron and fission damage. Annealing of this damage in UO2 occurs at temperatures
above 500  C, so lattice defects are thought to have
only a minor influence under normal operating
conditions. Similarly, a loss of effective conductivity due to pellet cracking may be annealed at higher
temperatures. It is not immediately clear whether
these annealing mechanisms hold similarly for
ThO2.
 A longer-term decrease in thermal conductivity
due to the formation of fission products. For
UO2, the thermal conductivity decreases at a rate
of 6–8% per 10 GWd per t UO2.11
 The modification of the porosity structure. The Xe
and Kr formed during irradiation may cause the
formation of large bubbles at higher burnups or
higher temperatures, lowering the overall thermal
conductivity.
An overview of information from irradiation tests
is given in Section 3.04.5.2.1–7. The discussion
mostly relates to (Th,U)O2, with a shorter piece on
(Th,Pu)O2 at the end (see Chapter 2.17, Thermal
Properties of Irradiated UO2 and MOX).
3.04.5.2.1 Restructuring

Short zero burnup ramp tests have been performed
in the NRX reactor on (Th,U)O2 with uranium
concentrations between 2.7% and 19% in order to
study the restructuring under fast reactor conditions
and establish the linear powers needed for central
fuel melting.82,83 It was found as expected that the
power ‘to melt’ decreased with uranium concentration in the range 5.4 –19%. After irradiation, pellets


that were sintered in an oxidative atmosphere
showed more grain growth than the ones sintered
in a reductive atmosphere. An irradiation of fuel
containing up to 50% uranium was carried out to
study grain growth in more detail.84 Columnar grain
growth appeared similar to that in UO2 fuel, but
proceeded at an estimated 350  K higher than for
UO2. In two other studies, it was estimated that
about 10–20% more power was needed to produce
a given structural change for thoria-based fuel,85 in
line with the idea that thoria offers more dimensional stability.
3.04.5.2.2 Thermal conductivity

A low-burnup linear power ramp test at a uranium
concentration of 10%86 has indicated that the in-pile
thermal conductivity of (Th,U)O2 follows that of pure
UO2 until about 1000  C, above which point UO2
conductivity becomes independent of temperature,
but (Th,U)O2 conductivity decreases further. This
trend seems in line with the presence of an electronic
contribution for UO2,40 but the observed crossover
temperature is much too low. A qualitative explanation is that annealing of irradiation damage is much
larger for UO2 (Section 3.04.3.3). Very limited data
are available regarding the extent of short-term thermal conductivity loss in ThO2- based fuel, and the
related annealing temperature. In a low-temperature
measurement of the in-pile thermal conductivity
of (Th,U)O2 containing 1.3% UO2,87 the initial conductivity of 12.5 W mKÀ1 at 60  C decreased rapidly
with increasing exposure, to a 50% lower plateau
value; postirradiation annealing led to a virtually

complete recovery at 1000  C. On the other hand,
Jacobs compared the unirradiated and in-pile irradiated conductivity of Th0.9U0.1O2 and concluded
that there was no statistical difference.11 Matolich
and Storhok measured the postirradiation thermal
conductivities of ThO2–UO2 pellets containing 3,
10, and 15% uranium, which had been irradiated
under varying conditions.11 No significant differences were observed for the 3% sample, which was
attributed to the relatively high irradiation temperature (900  C on average). The 10% sample showed
a higher postirradiation conductivity, which was
explained as due to the formation of columnar grains.
The higher burnup 15% sample finally showed a loss
of thermal conductivity of more than 50% even after
annealing at 1200  C. Taken together, these observations do not provide clear conclusions, but reveal
the many different parameters that influence the
thermal conductivity.


Thorium Oxide Fuel

3.04.5.2.3 CANDU fuel

The overall conclusion from a summary of Canadian
experience from the 1960s88 was that thorium-based
mixed oxide behaves similarly to UO2 when compared at the same fraction of the melting temperature.88 North American experience from the early
1970s has been summarized in contributions for the
ANS winter meeting 1977.80,89,90 From a study on
thorium–uranium oxide fuel containing 1.6% uranium in a CANDU-type reactor (linear power up
to 650 W/cm, burnup 16.7 GWd per tHM), it was
concluded that ‘FGR is approximately 1/10 that
of UO2 fuel operating at the same power.’89 Other

conclusions from CANDU studies on thoria fuel containing 1–3% uranium are that power ramping after
prolonged irradiation at lower powers induces defects
consistent with UO2 fuel, but that ‘fission product
release rates (following cladding failure) are between
one and two orders of magnitude lower, and fuel
deterioration is much less.89 A series of irradiations
in the NRX (heavy water-moderated) reactor led to
the conclusion that for higher uranium concentrations of 5–10%, the (Th,U)O2 fuel behaved ‘essentially the same as UO2 fuel,’ whereas even higher
concentrations (up to 19%) ‘resulted in poorer
performance.’90
An extensive set of Canadian (CANDU) highburnup irradiations up to 1985 is summarized in
Hastings et al.91 Nine irradiations are described of
pinlets and multipin (Th,U)O2 fuel bundles containing pellets produced with different fabrication parameters. In these experiments, the maximum linear
power was 570 W cmÀ1 and maximum burnup around
35 GWd per t. No failures were observed. Fission gas
release was never above 5%. Intermediate power
ramping up to 600 W/cm at a maximum burnup of
21 GWd per t in the BDL-421 experiment did not
cause any failure. It should be stressed here that
CANDU fuel contains little fissile material, and
because HEU was used, the investigated uranium
concentration was only 1–3%.
3.04.5.2.4 United States

American experience from four extensive irradiation
programs is discussed in detail in Hart et al.,11 which
forms a very useful resource for data on thorium oxide
fuel irradiations. Here we summarize the most important findings, leaving out the experimental details.
(Th,U)O2 fuel with U concentrations of 6.4%,
12.7%, and 25.6% were irradiated in the MTR and

the engineering test reactor (ETR) at very high linear
heat rates more typical for fast reactors. High fission

101

gas release was observed for especially the 12.7% and
25.6% samples. Postirradiation examination showed
that significant columnar recrystallization was accompanied by oxygen release; the decrease in thermal
conductivity associated with the oxygen release
served to explain the high fission gas fractional release.
During the Thorium Utilization program, sol–gel
sphere-pac (Th,U)O2 fuel (4.5% U) was compared
with pellet fuel, at high burnups under LWR conditions. The behavior of the different fuel types was
similar, and both types ‘had basic performance characteristics for a power reactor fuel up to 10 at.%
burnup.’11 Fission gas fractional release was generally
less than 20%, and fuel volume changes were less than
0.5% per % burnup. In another high-burnup irradiation on pellet fuel of the same composition and a
burnup of 10%, fission gas fractional release was
14%, ‘comparable to other oxide fuels under similar
conditions.’
During the Indian Point PWR program, a full core
of (Th,U)O2 pellet fuel containing 6.5 and 9.1%
U operated at linear powers of 70–410 W cmÀ1 to
burnups of 3 at.% with little change except for fuel
cracking. Fission gas fractional release was below 2%
in all cases, and it was pointed out that a similar
amount was expected from UO2 fuel under similar
conditions.
The extensive experience from the Shippingport light water breeder reactor program has
been summarized in Atherton et al.92 and Olson

et al.93 The performance of the seed and blanket
ThO2/233UO2 fuel rods was deemed ‘excellent,’
with no conclusive evidence of failure (the ‘seed’
fuel had a uranium content of 5–6 wt%, whereas
the blanket fuel contained only 1.5–3 wt%). Notable conclusions from postirradiation examination
were as follows92:
 Fission gas fractional release at the end of the core
life was less than 0.2%
 The lack of significant grain growth confirmed that
the operating temperature was below 1415  C and
that the fuel was stable.
 Fuel pellets remained cylindrical and essentially
intact. Limited axial and circumferential cracks
were observed, but there were no fuel chips or
dislocated pellet fragments.
 Fuel pores remained small up to burnups of
20 GWd per tHM; larger grain boundary fission gas
bubbles became evident only at higher burnups.
 No oxide was observed on the fuel side of the
zircaloy cladding.


102

Thorium Oxide Fuel

 Essentially no iodine was detected in the fuel–
cladding gap or on the cladding inner surface, but
low levels of 137Cs were present.
From Shippingport irradiation tests on fuel with

a wider range of uranium concentrations (0–20%),
fission gas release measurements irradiated at peak
linear powers from 72–738 W/cm up to burnups in
the range 0.9–56.5 GWd per tHM have yielded release
rates in the range 0.1–5.2% with an average below
1%, ‘much less than typical for UO2 fuels.’80 The
stability of fuel with up to 10% U in the temperature
range 600–1000  C was superior to that of UO2 at the
evaluated burnups (<1%),11 whereas thermal conductivity was unchanged from fresh fuel for a sample
with a burnup of 0.03%. Low-temperature creep
studies, however, indicated that fission-induced
creep was higher than in UO2 fuel. For samples containing less uranium (2 instead of 10%), the creep was
again higher by 25%. The following general conclusions from US experience are given in Hart et al.11:
1. Fission gas release for fuel with uranium content
up to 10% is comparable to that of UO2 under
typical LWR conditions. Within the investigated
range, increasing the UO2 content has little effect
on the release behavior.
2. Dimensional stability is better than for UO2, and
very little irradiation-induced swelling occurs. Up
to 10% burnup, the volume change is less than 1%
per % burnup.
3. Thermal conductivity remains unchanged under
irradiation but decreases with increasing UO2
content. Thermal expansion behavior is essentially
the same as for UO2 fuel (note the disagreement
with perhaps more accurate out-of-pile measurements, Section 3.04.3.2).
4. Structural changes under irradiation are generally
less than for UO2 under similar circumstances.
5. Due to the similarities in behavior, a fast route to

the licensing of thoria-based fuels would be to
establish correlations with the existing database
on UO2 fuel under normal operating conditions.
Normalization of the UO2 database for ThO2based fuel would require significantly less datapoints. However, off-normal and safety-related
data will be necessary.
3.04.5.2.5 India

In India, ThO2 rods have been tested in the central
parts of some reactor cores as a means to flatten
the power profile.94 From a ThO2 fuel element of
the Kakrapar power station with a burnup of 11 GWd

per tHM, no release was observed.95 Finally, Shiba
et al. found that xenon release from (Th,U)O2
powders increased with urania content and that the
release was more or less independent of preparation
conditions.96
3.04.5.2.6 Fission product behavior

(See Chapter 2.20, Fission Product Chemistry in
Oxide Fuels) Postirradiation fuel composition analyses of Th0.8U0.2O2 by Berman97 and Th0.95O0.05O2
by Padden et al.98 revealed the presence of metallic
globules of Mo alloyed with Tc,Ru,Rh,Pd at the edges
of columnar grains. These inclusions are also present
in high-burnup UO2 and (Pu,U)O2 fuel. ‘Gray phase’
perovskite inclusions of the BaZrO3 type, however,
were not observed. Ugajin et al. studied the chemical
form of solid fission products in some detail by reproducing the elemental composition of high-burnup
Th0.81U0.19O299 and did find the expected ‘gray
phase’ inclusions in the form of (Ba,Sr)(Zr,Ce)O2.

Compared with (U,Pu)O2 fuel, the absence of uranate
and plutonate and the large enrichment of Ce in the
gray phase are notable; Cs was also absent from
the gray phase. The solutes (rare earths, Ce, Zr)
were found to induce a contraction of the unit cell of
the fluorite matrix. The authors finally argue that the
thermochemistry of the solid fission products in pure
ThO2 would be similar.99 Momin et al. quantitatively
evaluated the unit cell contraction induced by the
dissolution of rare earths and confirmed that, unlike
UO2, ThO2 and (Th0.8U0.2)O2 did not form singlephase fcc solutions of Ba and Sr at all.100
Some experimental results on the diffusion of
volatile fission products in thorium oxide fuels
(I (Br), Cs (Rb), Te) reveal the same trends that
are indicated by fission gas release measurements
(Xe and Kr): a drastic slowdown of diffusion is
observed as compared with urania,101–105 to the
point that bulk diffusion of I and Te was ‘too low to
be measurable’ below 1000  C for trace-irradiated
Th0.98U0.02O2.102 Initially, the activation energy of
migration increases and the release decreases with
burnup, but these changes level off at higher burnups.104 (Th,U)O2 fuel containing 10% uranium has
been irradiated in the the experimental boiling water
reactor (EBWR) up to a uranium burnup of 3% while
monitoring the release of volatile species.106 Escape
of radionuclides, notably iodine, was ‘much smaller
than for UO2 fuel in PWRs,’ while only Xe and Kr
were significantly released. Similarly, tests with defected fuel in the EBWR showed no increase in gross
activity of the reactor water.11



Thorium Oxide Fuel

conductivity was observed at lower temperatures up
to at least 850  C, while at 1000  C, the thermal
conductivity was unchanged from the fresh fuel.
This indicates that the annealing temperature is
roughly 350–450  C above that of UO2, in line with
the observations for (Th,U)O2.
In Europe, the Oxide fuels, Microstructure and
Composition variations (OMICO) and Thorium
Cycle projects have been conducted in the past
decade. OMICO109 included the validation of a
(Th,Pu)O2 module in LWR fuel codes by a fuel irradiation in the BR-2 reactor in Belgium. Two irradiations
have been performed within the Thorium Cycle
project,19 which included a physicochemical study
on thorium-based oxides.20 One fuel pin containing
(Th,Pu)O2 with a fissile Pu content of 3% has been
irradiated in the KWO PWR in Obrigheim, Germany,
under typical PWR (U,Pu)O2 conditions (maximum
linear power 194 W cmÀ1) to a burnup of 37.7 GWd per
tHM (Figure 6). The dimensional stability of the fuel
was found to be ‘similar to that of PWR UO2 fuel.’110
In addition, (Th,Pu)O2 containing 11% Pu was
compared with regular (U,Pu)O2 (10% Pu) and
11% enriched UO2 in the high flux material testing
reactor (HFR) in Petten (NL) at maximum linear
powers of 200 W cmÀ1 up to a burnup of 39–
49 GWd per tHM. The fuels reached a burnup
of 45 GWd per tHM without the occurrence of

detrimental effects, confirming the suitability of
thorium-based mixed oxide as a light water reactor
matrix material for Pu burning that is capable of
reaching reasonable burnups. Radial and axial
swelling were observed to be the lowest for the
enriched UO2 fuel, whereas the swelling of
(Th,Pu)O2 was found to be slightly less than that

3.04.5.2.7 (Th,Pu)O2

Experience with (Th,Pu)O2 fuels has been building
more slowly. In India, much experience has been
gained through a series of irradiations in the CIRUS
reactor.8 A 6-pin cluster containing 5 ThO2–4%
PuO2 fuel and a dummy pin has been irradiated in
a pressurized water loop, at a maximum linear power
of 280 W cmÀ1 and up to a burnup of 18.5 GWd
per t.8,77,107 No failures occurred, and the performance was deemed ‘satisfactory.’
In Canada, 6 (Th,Pu)O2 fuel bundles have been
irradiated in the NRU reactor to burnups in the range
19–49 GWd per tHM and at peak powers of 520–
730 W/cm.108 Fission gas release rates of 1–5% were
obtained or burnups up to 36 GWd per tHM, significantly lower than those observed for UO2 and (U,Pu)
O2 CANDU fuel with similar power history. The
results were attributed to the relatively low temperature in the fuel in comparison with UO2. Hastings
et al.91 mention one additional Canadian experiment
on (Th,Pu)O2 fuel with a Pu concentration of 1.75%.
No problems were reported early in the irradiation,
at a maximum burnup of close to 5 GWd per t and
linear power up to 600 W cmÀ1. The in-pile thermal

conductivity of (Th,Pu)O2 containing low amounts of
Pu (1–3%) has been measured at low burnup up
to a maximum central temperature of 1650  C.85
The radially integrated thermal conductivity was
found to be better than that of UO2. This difference
increased with temperature so that for the maximum
central temperature, the thermal conductivity
was about 25% lower than that of unirradiated
ThO2, and about 25% higher than that of UO2 irradiated under similar conditions. As with (Th,U)O2,
irradiation-induced reduction of the thermal

(a)

103

(b)

1 mm

Figure 6 (a) Axial ceramograph of an unirradiated (Th,Pu)O2 pellet containing 3% Pu. (b) Radial ceramograph taken
after irradiation in the KWO Obrigheim reactor. Reproduced from Somers, J.; Papaoiannou, D.; Sommer, D. Irradiation of
thorium–plutonium mixed oxide fuel to 37.7 GWd/t in the Obrigheim pressurised water reacter (KWO). In Proceedings
of GLOBAL 2009, Paris, France, 2009.


104

Thorium Oxide Fuel

of regular (U,Pu)O2 fuel, when taking into account

the differences in burnup.111

3.04.6 Reprocessing and
Refabrication
Reprocessing (see Chapter 5.14, Spent Fuel Dissolution and Reprocessing Processes and Chapter
5.15, Degradation Issues in Aqueous Reprocessing
Systems) of spent UO2 and (U,Pu)O2 fuel for LWRs
(i.e., the PUREX process) is currently performed in
France and Japan and consists of the following steps:
 Head end: Mechanical processing (‘chopping’) and
dissolution of spent fuel. UO2 and (U,Pu)O2 (up to
plutonium contents of 35%112) may be dissolved in
concentrated nitric acid, allowing for separation
from the insoluble cladding.
 Separation of U, Pu, and fission products into
different stream solutions, based on the partitioning of the different components over aqueous and
organic phases.
 Purification of the different product and waste
streams.
 Back end: Vitrification of waste and conditioning of
the reusable U (as UO2(NO3)2) and Pu (as PuO2).
A representation of the separation steps is given in
Figure 7(a). In short, U and Pu are first extracted
from the nitric acid solution into an organic phase
containing 30% tri-n-butylphosphate (TBP), which
has a strong affinity for ions of valence þ4 and þ6
and thus separates Pu(IV) and U(VI) from most of the
fission products. The Pu(IV) is then reduced to
Pu(III) with a much lower solubility in the organic


phase, after which it is back-extracted (‘stripped’) into
an aqueous phase. The now roughly separated uranium and plutonium streams are decontaminated
by additional extraction–back-extraction cycles, in
which the TBP concentration, the acid concentrations in aqueous and organic phases, and the temperature are varied to obtain desired separation levels.
Although the technical feasibility of reprocessing
thorium–uranium mixed oxide was demonstrated in
the 1950s in the United States,5,113–116 where relatively large reprocessing plants based on the
THOREX reprocessing scheme have operated for
years,117,118 and later in Germany,119 thorium reprocessing has not fully matured. The main reasons are
the ban on using the HEU needed to seed the thorium, and the industries’ choice for the uranium–
plutonium fuel cycle. Issues specific to thorium,
reviewed in references,117,120,121 are discussed in the
next subsection.
3.04.6.1

The THOREX Process

The THOREX process was developed to separate
fissile uranium from the fertile thorium matrix, in
analogy with the PUREX process. The high stability
of the ThO2 matrix, which is an advantage when
considering in-pile behavior and waste disposal, renders ThO2 and (Th,U)O2 practically insoluble in
nitric acid. Spent thorium fuel is instead dissolved
using the ‘THOREX reagent’ (13 M HNO3 þ 0.03–
0.05 M FÀ þ 0.1 M Al3þ).5,119 Here, the fluoride (as
HF) is added to the nitric acid solution in order to
catalyze dissolution. However, the fluoride also facilitates corrosion of the stainless steel reactor vessels.

PUREX
FP (aq)


Pu(III) (aq)

U, Pu(IV), FP (aq)

(org)

U, Pu(IV) (org)

U (org)

(a)
THOREX
FP (aq)

Th (aq)

Th, U, FP (aq)

(org)

Th, U (org)

U (org)

(b)
Figure 7 Overview of the separation stage in spent fuel reprocessing, showing similarities between PUREX and
THOREX. (a) U–Pu separation in PUREX and (b) Th–U separation in THOREX.



Thorium Oxide Fuel

105

100
Partition coefficient

10

U(VI)
Th(IV)

1

Pu(IV)

Zr

Nb

Ru

0.1

Pu(III)

0.01

0


2

4

Acidity (M)

6

8

10

12

14

Figure 8 Partition coefficients (ratios of concentrations in the organic and aqueous phase, respectively) of Th, U, and Pu, as
well as for some more problematic fission products. The numbers are for extraction by 20% TBP in kerosene. Merz, E.
Wiederaufarbeitung im Thoriumbrennstoffkreislauf: ein Problemkatalog, KfA Ju¨lich, 1984.

Aluminum ions are added (as Al(NO3)3) to reduce
this problem by forming a complex with the fluoride.
Dissolution of the fuel may be further enhanced
by starting from a lower sintered density compared
with that of UO2-based fuel. In addition, a reduction
in the dissolution time by a factor 2 may be obtained
by adding about 1.5% MgO to the fuel (which also
enhances sintering of the pellets;3 Section 3.04.4.1).
It is to be noted that the zircaloy cladding was found
to be sufficiently resistant to corrosion by the ‘thorium reagent.’116

Figure 7(b) gives an overview of the THOREX
separation phase. As for PUREX, the first step
involves extracting the heavy metals from the aqueous phase by complexation with TBP:
Â
Ã
Th4þ þ 4NOÀ
3 þ xTBP < À > ThðNO3 Þ4 xTBP
Figure 8 gives partition coefficients (ratios of concentrations in the organic and aqueous phase) for the
relevant actinides. It may be clear from the figure that
Th(IV) is less extractable than U(VI) or Pu(IV). This
necessitates the use of a salting agent119 to drive the
complexation reaction to the right-hand side by
increasing the concentration of NOÀ
3 counterions.
Al(NO3)2 has been used for this, but it has the disadvantage that the aluminum adds to the volume of the

waste stream. The most widely used salting agent is
therefore HNO3. This case is referred to as the ‘acid
THOREX’ flowsheet.
To avoid the formation of a second organic phase,
the concentrations of thorium and HNO3 should be
limited; the loading of thorium in particular is limited
to about 30% of the theoretical capacity. These concentration limits lead, however, to high thorium
losses during the extraction step (Figure 8). As a
countermeasure, concentrated HNO3 is added at
the end of the extraction step, where thorium concentrations are low.
The subsequent separation of Th and U rests on
the difference in extractability between U(VI) and
Th(IV). Here the absence of a reduction step is an
advantage with respect to the PUREX process, but

the smaller difference in extractability between the
heavy metal ions forms a drawback. In India, an
alternative route is being developed in which U and
Th are separated during the first extraction step. In
this case, less (5%) TBP is added to the organic
phase, resulting in U-only extraction.
3.04.6.2

Beyond Thorex

The THOREX flowsheets were developed in the
1950s and 1960s for fuels in which HEU contributes
the fissile material. However, with the current ban on


106

Thorium Oxide Fuel

HEU and with the aim of reducing plutonium stocks,
the more likely fuel to be used in LWRs is thorium–
plutonium mixed oxide. After reaching the required
burnup, this fuel contains significant amounts of Th,
U, and Pu. No scheme currently exists to separate the
three components, but there are some theoretical ideas
on what such a scheme would look like, and India has
concrete plans to breed 233U from Pu–Th mixed oxide
fuel in the currently developed advance heavy water
reactor. Wilson discusses the difficulties of full threeway
separation, which are related to interferences between

the requirements for PUREX and THOREX separation, and considers keeping the U and Pu streams
unseparated to simplify the flowsheets.120
3.04.6.3 Radiation Issues in
Reprocessing and Refabrication
In addition to described chemical issues, the
THOREX process presents some problems related
to radiation levels. First, the almost 100% fissile
nature of the uranium stream raises criticality issues,
which may be solved by changing the shape and size of
the reaction vessels and diluting the uranium stream.
Furthermore, the 233Pa precursor to 233U (Section
3.04.5) has a half-life of 27 days, which means that a
9 month cooling time (about ten half-lives) is necessary to avoid an increase in reactivity during reprocessing and refabrication. Cooling is in any case needed
due to the relatively large decay heat of thoriumbased fuel and from the perspective of recovering as
much 233U as possible.117
Furthermore, it has been mentioned in the introduction that irradiated thorium-based fuel carries
some 232U, which has daughters that emit hard
g-rays. This complicates handling during fabrication
and reprocessing considerably through the requirement for biological shielding. However, because the
initial step (232U to 228Th, t1/2 ¼ 73.6 years) is relatively slow, handling of freshly separated uranium
may be less problematic during the first year after
irradiation.2,118 Here we hit a trade-off with the
decay time of 233Pa.
After reprocessing, the thorium stream itself contains 228Th (t1/2 ¼ 1.9 years), which also leaves the
thorium output radioactive for decades. As a result,
it may be advantageous to blend the reprocessed
uranium oxide with fresh ThO2 and consider the
spent thoria as waste. In fact, due to its (previously
discussed) high chemical stability, and in particular
its high corrosion resistance, ThO2 is an excellent

waste form.75,122,123

3.04.7 Conclusions
From the above, the general conclusion can be drawn
that (Th,U)O2 fuel significantly outperforms enriched
UO2 for uranium concentrations up to about 5%,
while fission gas retention and dimensional stability
are still slightly better for concentrations up to about
15%.90 Much confidence has been built for the use of
thoria-based fuels in thermal reactors under accepted
UO2 operating conditions, for U concentrations smaller than 10%; the more limited experience with
(Th,Pu)O2 indicates very similar behavior. For U or
Pu concentrations above 15%, the limited data suggest that performance falls below that of UO2 on
account of significantly reduced thermal conductivity
and melting point depression. As the main interest in
ThO2-based fuel at the time of writing is in the
burning of Pu in LWRs or HWRs, further irradiation
testing of (Th,Pu)O2 in the 6–10% Pu concentration
range and to high plutonium burnups is warranted.
Fabrication procedures for thorium mixed oxides are
straightforward and analogous to that of UO2 and
(U,Pu)O2. The main technical challenges therefore lie
in the reprocessing steps, and subsequent (shielded)
refabrication. Without closure of the thoriumuranium cycle, its added value with respect to uranium resource savings will remain limited.

References
1. Clayton, J. C. American Chemical Society Meeting;
Pittsburgh, PA, 1993, />energycitations/servlets/purl/10191380-PWWUeT/
native/10191380.pdf.
2. Lung, M. Nuclear Science and Technology project report

EUR 17771 for the European Commission, 1997.
3. Anantharaman, K.; Shivakumar, V.; Saha, D. J. Nucl.
Mater. 2008, 383, 119–121.
4. Gruppelaar, H.; Schapira, J. P. Nuclear Science and
Technology project report EUR 19142 for the European
Commission, 2000.
5. The use of Thorium in Nuclear Reactors, document
WASH-1097 of the US Atomic Energy Commission, 1969.
6. Thorium based nuclear fuel: Current status and
perspectives, IAEA TECDOC 412, 1987.
7. Thorium based fuel options for the generation of
electricity: developments in the 1990s, IAEA TECDOC
1155, 2000.
8. Thorium fuel utilization: Options and trends, IAEA
TECDOC 1319, 2002.
9. Potential of thorium based fuel cycles to constrain
plutonium and reduce long lived waste toxicity, IAEA
TECDOC 1349, 2003.
10. Thorium fuel cycle – potential benefits and challenges,
IAEA TECDOC 1450, 2005.
11. Hart, M. E.; Griffith, C. W.; et al. ThO2-based pellet
fuels – their properties, methods of fabrication and
irradiation performance; DOE report PNL-3064; US, 1979.


Thorium Oxide Fuel
12.
13.

14.

15.

16.

17.

18.
19.

20.
21.

22.
23.

24.

25.
26.

27.
28.
29.
30.
31.
32.
33.
34.

35.

36.
37.

38.
39.
40.

Rodriguez, P.; Sundaram, C. V. J. Nucl. Mater. 1981, 100,
227–249.
Ganguly, C. Nuclear reactor fuel fabrication.
In Encyclopedia of Materials: Science and Technology,
2008; pp. 6324–6339.
Rodriguez, P. Thorium fuel cycles. In Encyclopedia of
Materials: Science and Technology, 2008; pp. 1–8.
Ramanujam, A. PUREX and THOREX processes.
In Encyclopedia of Materials: Science and Technology,
2008; pp. 7918–7924.
Uranium Red Book (Uranium 2009: Resources,
Production and Demand’), NEA/IAEA joint publication,
NEA No. 6891, 2009.
Herring, J. S.; MacDonald, P. E.; Weaver, K. D.;
Kullberg, C. Nucl. Eng. Des. 2001, 203(2001), 65–68,
1–20.
Magill, J.; Galy, J. Radioactivity Radionuclides Radiation;
Springer Verlag: Germany, 2005.
Schram, R. P. C.; Kuijper, J. C.; Sommer, D.; et al. The
thorium cycle project. In Proceedings of the International
Workshop on P&T and ADS Development, 2003.
Bakker, K.; Cordfunke, E. H. P.; Konings, R. J. M.;
Schram, R. P. C. J. Nucl. Mater. 1997, 250, 1–12.

Thermophysical properties database of materials for light
water reactors and heavy water reactors, IAEA Technical
Document 1496, 2006
Sobolev, V.; Lemehov, S. J. Nucl. Mater. 2006, 352,
300–308.
De Bruycker, F.; Boboridis, K.; Rini, M.; Konings, R. J. M.;
Somers, J.; Manara, D. The melting behavior of
plutonium dioxide. In Proceedings of the EMRS Spring
Meeting 2010, Strasbourg, France, 2010.
De Bruycker, F.; Boboridis, K. ;; Manara, D.; Po¨ml, P.;
Rini, M.; Konings, R. J. M. Materials Today 2010, 13(11),
52–55.
Carbajo, J. J.; Yoder, G. L.; Popov, S. G.; Ivanov, V. K.
J. Nucl. Mater. 2001, 299, 181–198.
Cordfunke, E. H. P.; Konings, R. J. M. Eds.
Thermochemical Data for Reactor Materials and Fission
Products; Elsevier, 1990.
Freshly, M. D.; Mattys, H. M. General Electric Report
HW-76559 1962, 11, 6.
Lambertson, W. A.; Mueller, M. H.; Gunzel, F. H. J. Am.
Ceram. Soc. 1953, 36, 397.
Slowinski, E.; Elliott, N. Acta Crystall. 1952, 5, 768.
Gmelins Handbuch der anorganischen Chemie, 8th
edition, Thorium, Springer Verlag, Berlin (vol. C1, C2).
Hubert, S.; Purans, J.; Heisbourg, G.; Moisy, P.;
Dacheux, N. Inorg. Chem. 2006, 45, 3887–3894.
Aitken, E. A.; Edwards, J. A.; Joseph, R. A. J. Phys.
Chem. 1966, 70, 1085.
Alexander, C. A.; Ogden, J. S.; Cunningham, W. Batelle
Memorial Institute Report BMI-1789; 1967.

Touloukian, Y. S.; Kirby, R. K.; Taylor, R. E.; Lee, T. Y. R.
Thermal Expansion Nonmetallic Solids; IFI/Plenum: New
York, 1970.
Olander, D. R. Fundamental Aspects of Nuclear Reactor
Fuel Elements, ERDA, 1976; Chapter 10.
Pillai, C. G. S.; George, A. M. J. Nucl. Mater. 2000, 277,
116–119.
Cozzo, C.; Staicu, D.; Somers, J.; Fernandez, A.;
Naisse, F.; Konings, R. J. Nucl. Mater. 2011, 416,
135–141; Proceedings of the 2010 EMRS Spring
Meeting.
Ronchi, C.; Hiernaut, J. P. J. Alloys Comp. 1996, 240,
179.
Fink, J. K.; Chasanov, M. G.; Leibowitz, L. Argonne
National Laboratory Report ANL-CEN-RSD-77–1; 1977.
Young, R. A. J. Nucl. Mater. 1979, 87(2–3), 283–296.

41.
42.
43.

44.
45.

46.
47.
48.

49.
50.

51.
52.
53.
54.
55.
56.
57.
58.
59.
60.

61.
62.
63.
64.
65.
66.
67.
68.
69.

70.
71.
72.

73.
74.
75.

107


Gibby, R. L. J. Nucl. Mater. 1971, 38, 163–177.
Golovnin, I. S. Atom. Energy 2000, 89, 627–637.
Springer, J. R.; Aldridge, E. A.; Goodyear, M. U.;
Wright, T. R.; Lagedrost, J. F. Fabrication,
characterization and thermal property measurements of
ThO2-UO2 fuel materials, BMI-X-10210; Oct 1967.
Murabayashi, M.; Nucl, J. Sci. Technol. 1970, 7, 559.
Springer, J. R.; Lagedrost, J. F. Thermal properties of
ThO2–UO2 fuel materials (LWBR development program),
Batelle Memorial Institute report BMI-X-10231; 1968.
Berman, R. M.; Tully, T. S.; Belle, J.; Goldberg, I. DOE
report WAPD-TM-908; 1972.
Basak, U.; Sengupta, A. K.; Ganguly, C. J. Mat. Sci. Lett.
1989, 8, 449–450.
Cox, J. D.; Wagman, D. D.; Medvedev, V. A. CODATA
key values for thermodynamics, Hemisphere: New York,
1989.
Martin, D. G. J. Nucl. Mater. 1988, 152, 94.
Osborne, D. W. J. Chem. Phys. 1953, 21, 1884.
Fischer, D. F.; Fink, J. K.; Leibowitz, L.; Belle, J. J. Nucl.
Mater. 2008, 118(2–3), 342–348.
Dash, S.; Parida, S. C.; Singh, Z.; Sen, B. K.;
Venugopal, V. J. Nucl. Mater. 2009, 393, 267–281.
Schram, R. P. C. J. Nucl. Mater. 2005, 344, 223–229.
Cohen, I.; Berman, R. M. J. Nucl. Mater. 1966, 18,
77–107.
Ugajin, M. J. Nucl. Mater. 1982, 110, 140.
Ugajin, M.; Shiratori, T.; Shiba, K. J. Nucl. Mater. 1983,
116, 172.

Ugajin, M. J. Nucl. Sci. Technol. 1983, 20, 228.
Kutty, T. R. G.; Hegde, P. V.; Khan, K. B.; et al. J. Nucl.
Mater. 2002, 312, 224.
White, G. D.; Bray, L. A.; Hart, P. E. J. Nucl. Mater. 1981,
96(3), 305–313.
Smit-Groen, V. M.; Schram, R. P. C.; Somers, J.;
Contrad, R. HFR Fabrication Report Thorium Cycle,
document Th5-DOC-016 of the 5th framework
program Thorium Cycle of the European Commission
(2002).
Fuller, E. L., Jr.; Agron, P. A. J. Coll. Int. Sci. 1976, 57,
193–200.
Lords Smith, R. E. J. Nucl. Mater. 2004, 328, 215–224.
Gammage, R. B.; Fuller, E. L., Jr.; Holmes, H. F.; Coll, J.
J. Coll. Int. Sci. 1972, 38(1), 91–96.
Kutty, T. R. G.; Khan, K. B.; Achuthan, P. V.; et al. J. Nucl.
Mater. 2009, 389, 351–358.
Kurina, I. S.; Gudkov, L. S.; Rumyantsev, V. N. Atom.
Energy 2002, 12, 461.
OECD/NEA report. Trends in the nuclear fuel cycle:
Economic, environmental and social aspects, 2001.
Kutty, T. R. G.; Khan, K. B.; Somayajulu, P. S.; et al.
J. Nucl. Mater. 2008, 373, 299–308.
Kutty, T. R. G.; Kulkarni, R. V.; Sengupta, P.; et al. J. Nucl.
Mater. 2008, 373, 309–318.
Kutty, T. R. G.; Somayajulu, P. S.; Khan, K. B.; Kumar, A.;
Kamath, H. S. J. Nucl. Mater. 2009, 384,
303–310.
Ganguly, C. Bull. Mat. Sci. 1993, 16, 509–522.
Kumar, N. Nucl. Tech. 1991, 96, 169–177.

Vaidya, V. N. Sol–gel process for ceramic nuclear fuels,
BARC newsletter, />nl/2002/200203–01.pdf.
Vaidya, V. N. J. Sol Gel Sci. Tech. 2008, 46,
369–381.
Pai, R. V.; Mukerjee, S. K.; Vaidya, V. N. J. Nucl. Mater.
2004, 325, 159–168.
Tables of Physical & Chemical Constants. 4.7.2 Neutron
cross-sections. Kaye & Laby Online, version 1.0
(2005) www.kayelaby.npl.co.uk


108
76.
77.
78.

79.

80.
81.
82.
83.
84.
85.
86.

87.

88.


89.
90.
91.

92.

93.

94.
95.

96.
97.
98.
99.
100.
101.

Thorium Oxide Fuel
Note that the Peach Bottom and Fort st. Vrain reactors
ran on thorium carbide fuel.
Schram, R. P. C.; Klaassen, F. C. Prog. Nucl. Energy
2007, 49, 617–622.
Fehrenbach, P. J.; Miller, A. I. In Nuclear Engineering
Handbook; Kok, K. D., Ed.; CRC Press, 2009;
Chapter 14.
Bailly, H. Menessier, D.; Prunier, C. Eds. The Nuclear
Fuel of Pressurized Water Reactors and Fast Reactors –
Design and Behaviour; CEA (Lavoisier Publishing):
France, 1999.

Goldberg, I.; Spahr, G. L.; Giovengo, J. F.; Waldman, L. A.
Trans. Am. Nucl. Soc. 1977, 27, 308–310.
Basu, M.; Mishra, R.; Bharadwaj, S. R.; Das, D. J. Nucl.
Mater. 2010, 403, 2204–215.
Rao, S. V. K. Investigation of ThO2-UO2 as a nuclear fuel,
AECL report 1785; 1963.
Rao, S. V. K. J. Nucl. Mater. 1964, 12(3), 323–329.
Neimark, L. A. Trans. Am. Nucl. Soc. 1962, 5(1).
Jeffs, A. T. Thermal conductivity of ThO2–PuO2 under
irradiation; AECL report 3294; 1969.
Kraerheim, G.; Rolstad, E. In-core study of fuel/clad
interaction and fuel centre temperature, OECD Halden
Reactor Project report HRP-107; 1969.
MacEwan, J. R.; Stoute, R. L. Annealing of irradiation
induced thermal conductivity changes in ThO2–1.3 wt%
UO2. In 70th Annual Meeting of American Ceramic
Society, Mar 1969; paper 11-N-68..
Lewis, W. B.; Duret, M. F.; Craig, D. S.; Veeder, J. R.;
Bain, A. S. Report AECL-3980; Atomic Energy of Canada
Limited, 1971.
Jones, R. W.; Lee, H. R.; Hahn, H.; Walter, J. F.; Celli, A.
Trans. Am. Nucl. Soc. 1977, 27, 304–305.
Lane, A. D.; Jones, R. W.; Hamel, D. Am. Nucl. Soc.
1977, 27, 305–306.
Hastings, I. J.; Carter, T. J.; MacGillivray, G.;
MacDonald, R. D.; Judah, J. Canadian High burnup
fuel experience. In Proceedings of 6th Annual
Conference of the Canadian Nuclear Society, 1985.
Atherton, R. Water cooled breeder program
summary report; DOE R&D report WAPD-TM-1600;

1987.
Olson, G. L.; McCardell, R. K.; Illum, D. B. Fuel
summary report: Shippingport Light Water Breeder
Reactor; INEEL report EXT-98–00799; 2002.
Anantharaman, K. Irradiation experience with thoria
based fuels, paragraph 3.2 in BARC report 1532; 1990.
Viswathan, U. K.; Anantharaman, S.; Sahoo, K. C.
Measurement of fission gas release from irradiated
nuclear fuel elements, Report BARC/2005/E/026 of the
Bhabha atomic research centre, 2005.
Shiba, K.; Itoh, A.; Ugajin, M. J. Nucl. Mater. 1981, 96(3),
255–260.
Berman, R. M. WAPD-TM-1236, 1976.
Padden, T. R.; Burton, R.; Campbell, C. E. WAPD-TM-644,
1976.
Ugajin, M.; Shiratori, T.; Shiba, K. J. Nucl. Mater. 1979,
84, 26–38.
Momin, A. C.; Mirza, E. B.; Mathews, M. D. J. Nucl.
Mater. 1991, 185, 308–310.
Das, D.; Kaimal, K. N. G. Release behavior of gaseous
and volatile fission products in irradiated thoria based
fuel materials – a laboratory scale study. In Conference
on Proceedings of High Burn-Up Issues in Nuclear Fuels,
Mumbai, India, 2005; pp. 137–142.

102.
103.
104.
105.


106.
107.

108.
109.

110.

111.
112.

113.
114.

115.
116.

117.

118.

119.

120.

121.

122.

123.

124.

Kaimal, K. N. G.; Kerkar, A. S.; Shirsat, A. N.; et al.
J. Nucl. Mater. 2003, 317, 189–194.
Shiba, K.; Itoh, A.; Akabori, M. J. Nucl. Mater. 1984,
126(1), 18–24.
Akabori, M.; Fukuda, K. J. Nucl. Mater. 1990, 186(1),
47–53.
Matzke, H.; Ray, I. L. F.; Verrall, R. A. Diffusion
and behaviour of iodine and rubidium in UO2 and ThO2,
183. In: Water reactor fuel behaviour and fission
product release in off-normal and accident conditions,
IAEA document IWGFPT/27, 1986.
Robertson, R. F. S. Tests on Defected thoria–urania
fuel specimens in EBWR, ANL report 6022; 1960.
Ganguly, C. ThO2-based fuels for PHWRs – Fabrication,
characterization and test irradiation. In 2nd International
Conference on CANDU Fuel, Canada, 1989, 398.
Karam, M.; Dimayuga, F.C.; Montin, J. Fission gas
release of (Th,Pu)O2 CANDU fuel.
Verwerft, M.; We´ber, M.; Lemehov, S.; et al. Oxide fuels:
microstructure and composition variations (OMICO).
In Proceedings of the FISA, 2003.
Somers, J.; Papaoiannou, D.; Sommer, D. Irradiation
of thorium–plutonium mixed oxide fuel to 37.7 GWd/t
in the Obrigheim pressurised water reacter (KWO).
Proceedings of GLOBAL 2009, Paris, France, 2009.
Hania, P. R. et al. unpublished results from the HFR
Thorium Cycle irradiation.
Treatment and recycling of spent nuclear fuel,

M. Lecomte (editor), a CEA Nuclear Energy Division
monograph.
Flag, J. F. Ed. Chemical Reprocessing of Reactor Fuels;
Academic Press, 1961.
Stoller, S. M.; Richards, R. B., Eds. Reactor Handbook,
Vol. II: Fuel Reprocessing; Interscience: New York, 1961;
2nd edn.
Blanco, R. E.; Ferris, L. M.; Ferguson, D. E. Aqueous
Processing of Thorium Fuels; ORNL-3219; 1962.
Nicholson, E. L.; et al. Recent developments in
thorium fuel processing. In Second Symposium on
Thorium Fuel Cycle, Gatlinburg, TN, USA, 1966.
Greneche, D.; Szymczak, W. J.; Buchheit, J. M.;
Delpech, M.; Vasile, A.; Golfier, H. Rethinking the thorium
fuel cycle: An industrial point of view. In Proceedings of
the ICAPP Conference; 2007; Paper 7367.
Brooksbank, R. E.; McDuffee, W. T.; Rainey, R. H.
A review of thorium reprocessing experience, ORNL
report for the DOE.
Zimmer, E.; Merz, E. Chemical processing of HTR fuels
applying either THOREX or PUREX flowsheets.
In Proceedings of the Specialists’ Meeting on
Gas-Cooled Reactor Fuel Development and Spent Fuel
Treatment, 1983, IAEA IWGGCR8.
Wilson, P. D. Development requirements in reprocessing
thorium fuels; In Thorium fuel utilization: options and
trends; IAEA TECDOC 1319; 2002 [5].
Carter, W. L. Thorium Utilization Program: A Survey of
Processing Methods for Thorium Reactor Fuels; Report
ORNL-TM-241; Oak Ridge, 1962.

Markowitz, J. M.; Clayton, J. C. Corrosion of oxide
nuclear fuels in high temperature water, WAPD-TM-909;
1970.
Thomas, L. E.; Einziger, R. E.; Buchanan, H. C. J. Nucl.
Mater. 1993, 20(1), 310–319.
Tel, H.; Eral, M.; Alta, Y. J. Nucl. Mater. 1998, 256,
18–24.