Comprehensive nuclear materials 3 04 thorium oxide fuel

Comprehensive nuclear materials 3 04 thorium oxide fuel Comprehensive nuclear materials 3 04 thorium oxide fuel Comprehensive nuclear materials 3 04 thorium oxide fuel Comprehensive nuclear materials 3 04 thorium oxide fuel Comprehensive nuclear materials 3 04 thorium oxide fuel Comprehensive nuclear materials 3 04 thorium oxide fuel Comprehensive nuclear materials 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.

3.04.2.1 Thorium as an Abundantly Available Resource for Nuclear Fuel 89

ANS American Nuclear Society

BWR Boiling Water Reactor

CANDU CANadian Deuterium Uranium reactor

FGR Fission Gas Release

HEU High Enriched Uranium ( >20% U-235)

HMTA Hexamethylene tetramine

IAEA International Atomic Energy Agency

HTR High Temperature Reactor

LEU Low Enriched Uranium ( <20% U-235) LWR Light Water Reactor

MSR Molten Salt Reactor

PUREX Plutonium Uranium EXtraction PWR Pressurized Water Reactor

TRISO Tristructural Isotropic

87

THOREX THORium EXtraction

XRD X Ray Diffraction

EBWR Experimental Boiling Water Reactor

ETR Engineering Test Reactor

MTR Materials testing reactor

3.04.1 Introduction

Thorium (Th) is an actinide element with atomic

number 90 It is a silvery-white colored metal,

discov-ered 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
1010

years In addition, other Th-isotopes are

found in nature in trace amounts, as

daughter-pro-ducts in the decay chains of uranium (227Th,230Th)

and thorium itself (228Th) These isotopes are much

shorter lived than their mother isotopes and,

conse-quently, much less abundant As232Th 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

neu-tron in 232Th, 233U is formed, according to the

following nuclear reaction:

232Thþ n !233Th!233Pa!233U

Protactinium (233Pa), 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

transmu-tation of238U into239Pu As thorium is more abundant

than uranium, with otherwise similar chemical and

(neutron) physical properties, it opens up the

possi-bility 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 of233U bred from232Th,

sufficient fissile material must be generated first

Initi-ally, the breeding of233U must be assisted by a

suffi-cient amount of fissile material in the core of the

nuclear reactor This can be done by either enriched

uranium (with enrichments higher than in

conven-tional fuel) or plutonium

The potential of thorium was identified in the early

stages of nuclear technology development The

Ship-pingport atomic power reactor in the Unites States was

fueled with thorium to establish a 232Th/233U fuelcycle, from 1977 until its decommissioning in 1982.1

In Germany, the THTR-300, a high-temperaturereactor (HTR), operated from 1983 to 1989 with acore consisting of HTR pebbles with thorium and(highly enriched) uranium

However, because uranium proved to be dantly available to cover the world demand for theproduction of nuclear fuel, the necessity to develop athorium fuel cycle on an industrial scale never arose.2Nevertheless, the potential of thorium as a resourcefor nuclear power has always been recognized andresearched, for example, in countries with large tho-rium reserves and less access to uranium resources.India is probably the best example of a country thathas developed a comprehensive approach to a sus-tainable nuclear fuel cycle based on thorium.3

abun-In the last decade of the twentieth century, asecond incentive came up, that of using thorium as

a way to reduce the radiotoxicity of spent fuel Thisradiotoxicity is dominated by the transuranium ele-ments plutonium, americium, and curium The use ofthorium instead of 238U as the main fertile isotopereduces the amount of transuranium isotopes in thespent 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 consid-ered was to use thorium as a means to efficientlyincinerate excess military (weapons-grade, WG) plu-tonium from the disposition programs in Russia andthe United States In fact, thorium is a suitable matrixfor plutonium

This chapter focuses on thorium oxide as nuclearfuel The oxide is the most investigated chemicalform and the one most likely to be used in the exist-ing reactor fleet (but to give other examples: ThZrmetal fuel has been developed in the United States,and thorium mixed carbides and oxicarbides havebeen studied in the United States and Germany;see Chapter3.02, Nitride FuelandChapter 3.03,Carbide Fuel)

Much of the relevant research from the last tury is available in numerous overview articles,2,4,5and the IAEA ‘Technical Documents’6–10; theseworks give a fairly complete picture but tend tofocus on reactor physics and then mostly on (Th,U)

cen-O2 fuels containing high enriched uranium (HEU).Two other useful sources are a comprehensiveBatelle report from 1979 assessing the available data

on properties, fabrication, and irradiation mance of ThO –UO pellet fuels11 and a review

perfor-article by Rodriguez and Sundaram from 1981

con-taining 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

proper-ties, 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

tho-rium as a fissile source material is highlighted In

Section 3.04.3, the basic physical properties of

tho-rium oxide fuels are discussed in comparison with the

oxides of uranium and plutonium The fabrication

aspects are discussed inSection 3.04.4 In Section

3.04.5, the behavior of thorium oxide fuel under

irradiation is treated Reprocessing issues are

dis-cussed inSection 3.04.6

The authors do not claim to have presented a

complete picture of thorium in every detail

How-ever, 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

incinera-tion of excess WG-Pu These three applicaincinera-tions 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

con-tent 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

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) aregiven in Table 1, taken from the Uranium Red Book

2009.16 This overview estimates a total of knownresources of 2.5 million tons Th (identified resources)and an additional prognosticated amount of 1.9million tons

Currently, there is little demand (and industrialneed) for thorium; it is obtained mostly as a by-product from the production of rare earth metals

or from the production of titanium-, zirconium-, ortin-bearing minerals from monazite deposits.16Littlesystematic exploration has therefore been conductedspecifically for thorium, which explains the relativeuncertainty about the total world reserves

The large availability and relatively easy bility have been a large incentive to look at thorium

retrieva-as nuclear fuel.17Specifically, this has been the case incountries with large thorium reserves, such as India,Brazil, and, recently, Norway

The largest advantages of 232Th over the parable fertile isotope 238U are the amount of neu-trons produced per fission and the fission-to-captureratio, which are both higher than in 238U (Section3.04.5.1) This allows breeding with thorium in ther-mal reactors as well, whereas for238U, a fast spectrum

com-is mandatory for breeding Thorium can thus be useddirectly in light or heavy water reactors or in HTR Itshould be noted, though, that achieving breedingwith thorium is not straightforward The exact gain

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

Country Identified Th resources

in resource efficiency, which can be achieved in a

thorium cycle, as compared with the ‘standard’

ura-nium 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 breeding233U together with the burnup

rate of the fissile topping, and the choice of whether

or not to reprocess the spent fuel, ultimately

deter-mines 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(seeChapter3.13,

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

ele-ments 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 UO2fuel 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

exten-sively in Europe under the 4th Framework program

‘Thorium as a waste management option’4and 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 isclear that the use of HEU offers the most benefits interms of radiotoxicity reduction But the use of HEU

in a civil nuclear fuel cycle is not preferred due toreasons of proliferation risks Furthermore, the effect

of resource efficiency through the use of abundantthorium is counterbalanced by the amount of naturaluranium material needed to enrich uranium toenrichments of 90% The combination of thoriumwith HEU, although preferred from a radiotoxicitypoint of view, is therefore not a viable option for asustainable fuel cycle The more the amount of ura-nium (238) or plutonium present in the initial fuel,the more the amount of transuranics produced, whichdetermines the lifetime of the spent fuel Neverthe-less, a significant reduction of the radiotoxicity can

be achieved Once an equilibrium thorium cycle isachieved, the benefits increase, as then the fuel

is based on the combination of232Th/233U only

3.04.2.3 Reduction of Excess MilitaryPlutonium

A third, more specific potential application of rium is connected to the incineration of excessmilitary plutonium More than 250 tons of WG-Pu,containing around 93%239Pu, has been produced inthe world for military purposes, mostly by the UnitedStates and the Russian Federation Part of these stock-piles has been declared as excess plutonium, and boththe United States and Russia have agreed to dispose of

tho-34 ton WG-Pu In the disposition program, it has beenwell defined how the excess plutonium will be incin-erated, that is, through its use as (U,Pu)O2 fuel innuclear power plants in the United States and as fuelfor fast reactors in Russia Nevertheless, the mixture

of thorium with WG-Pu into (Th,Pu)O2 fuel mayprovide a technical option to reduce the threat ofmilitary plutonium The application of (Th,Pu)O2fuel may similarly be used to more efficiently reducestockpiles of separated civil plutonium

3.04.3 Physical Properties of Thorium Oxide Fuel

The following section describes the importantthermophysical properties of ThO2 in comparisonwith UO2and PuO2 The comparison with the morecommon fuel oxides UO2and PuO2(seeChapter2.02,Thermodynamic and Thermophysical Properties

of the Actinide Oxides) is made throughout to show

that the three compounds are very similar and to

simul-taneously highlight the differences, which turn out

mostly favorably for ThO2 What sets ThO2fuel 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

nonstoichio-metry is reflected in a relatively high thermal

conduc-tivity (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 Crystal Structure

Close similarities exist between the three actinide

oxides: In common with UO2and PuO2, ThO2

pos-sesses an fcc lattice, and all three materials can be

heated to melt without undergoing phase

transi-tions.20In addition, thermal conductivities and

ther-mal 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 ThO2to pure PuO227or

UO2,28,29 the lattice parameter of the cubic lattices

changes linearly with the additive fraction (Vegard’sLaw behavior) However, locally the individual heavymetal–oxygen bond lengths in the mixed oxidestay somewhat closer to the values of the purecompounds.30,31

It follows from the Vegard’s Law behavior ofthe stoichiometric mixed oxide that their room-temperature densities may simply be obtained bylinear interpolation of the weights and cell volumesgiven inTable 2and additionally that the solid solu-tions thus formed show ideal behavior.20Vapor pres-sure measurements on ThO2–UO2 solid solutionshave indicated that the same ideality also holds athigher temperatures.32,33This allows one to reversethe argument and claim that the lattice parameteralso changes linearly with composition in the high-temperature region Therefore, the thermal expan-sion of the mixed actinide oxides can be obtained bylinearly interpolating the thermal expansions of thepure compounds.20

3.04.3.2 Thermal ExpansionTouloukian et al recommend the following relationfor the thermal expansion of pure ThO2

34based on

a large set of measurements in excellent agreement:DL=L0ð293 KÞ ¼ 0:179 þ 5:097
104T

þ 3:732
107T2

 7:594
1011T3ð150 2000 KÞThe thermal expansion coefficientaLis obtained bydifferentiation Figure 1 compares this expansionfor ThO2 with that observed for UO2, and PuO2,which have all been described using a single set ofrelations25:

Table 2 A comparison of physical parameters for the dioxides of Th, U, and Pu

Standard enthalpy of formation, DH0 J mol1 1226.4 1085.0 1055.8

a Value taken from a recent laser flash measurement, 124 but note that PuO2melts 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 Document 20

and thermophysical data from Bakker et al 19 and Cordfunke and Konings 23

DL=L0ð273 KÞ ¼ 0:266 þ 9:802
104T

 2:705
108T2

þ 4:391
1011T3ð273 923 KÞDL=L0ð273 KÞ ¼ 0:328 þ 1:179
103T

 2:429
107T2

þ 1:219
1010T3ð923 K TmÞwhere Tmis the melting temperature The compari-

son shows that thermal expansion is slightly lower for

ThO2 For a detailed discussion of

temperature-dependent theoretical density, linear expansion, and

melting points of ThO2and ThO2–UO2mixtures, 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

scat-tering 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 (4.20
104m K W1) and

B (2.25
104m W1) for 95% dense ThO2 More

recent experimental data are available from Pillai36

and Cozzo.37 The three results are in reasonable

agreement (Figure 2)

InFigure 2, a comparison is also made with thecorrelations given for UO225,38,39 and PuO2.37 Thethermal conductivity of UO2is obviously below that

of ThO2 in the temperature region shown in thisfigure (up to 1600 K); however, an electronic contri-bution to the thermal conductivity of UO2kicks in attemperatures above 2000 K, whereas this contribu-tion is absent for ThO2.40

For PuO2, Cozzo et al argue that previous ies41,42 yielding values close to those for UO2 hadmost probably been performed on samples of ill-defined stoichiometry and that stoichiometric PuO2

stud-in fact has the largest thermal conductivity of thethree oxides.37 This argument indicates that due tothe multivalence of the metal atoms, control over stoi-chiometry is not trivial Even when the stoichiometry iscarefully controlled during fabrication, PuO2and UO2will become nonstoichiometric upon heating or when

in oxidative or reductive environments This results,for many practical conditions, in some loss of thermalconductivity, which is difficult to control

Mixing of the actinide oxides generally depressesthe thermal conductivity, mostly because the additiveheavy metal ions act as phonon-scattering centers.The scattering term A is therefore affected morestrongly than the phonon interaction term B Gibbyhas observed this trend for the mixing of uranium andplutonium dioxide,41and on ThO2the depression isquite pronounced.43

Using a small dataset from Murabayashi44 andSpringer and Lagedrost45 selected from the availableliterature, Bakker et al.20derive values of the parameters

A and B for 95% dense Th1yUyO2and y up to 0.1(T ¼ 300–1800 K): A ¼ 4.195
104þ 1.112y  4.499y2(m K W1), B ¼ 2.248
1049.170
104y þ4.164
103y2 (m W1) These authors reject on

0

2 4 6 8 10 12 14

UO2, Lucuta-Ronchi, 95% TD

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

Figure 2 Thermal conductivity of stoichiometric ThO 2 ,

Figure 1 Thermal expansion of ThO 2 and UO 2 /PuO 2 In

the range shown, both relations are valid The expansion

curve for ThO 2 has been adjusted here to a reference

temperature of 273 K.

theoretical grounds an earlier assessment by Berman

et al.46from a partially overlapping dataset; however,

Berman gives correlations for higher uranium

con-centrations According to Bakker’s recommendation,

the addition of 10% UO2to ThO2results in a

reduc-tion of the conductivity by 40–50% More recently,

Pillai et al have found that the addition of as little as

2% UO2decreases the thermal conductivity of thoria

by 10–30%.36In the recent IAEA review,21new 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
103þ 0.572y  0.5937y2

(m K W1),

B ¼ 2.4
104(m W1) This correlation has a

mini-mum 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 0.08388 þ 1.7378
104y

(m K W1), B ¼ 2.62524 þ 1.7405
104y (m W1)

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)O2and (Th,Pu)O2fuels reveals significantly

different results, which seems to indicate measurement

uncertainties related to sample microstructure and

stoichiometry When using the Berman40and Bakker20

correlations for (Th,U)O2and the Cozzo37correlation

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 Document21suggest that the thermal

con-ductivity of UO2is approached with the addition of

about 20% uranium or only 6% Pu

3.04.3.4 Thermophysical Properties

The standard enthalpy of formation DfH0(298.15)

is1226.4 kJ mol1(Table 1),48which makes ThO2

the most stable oxide known As shown inTable 2,

this is reflected in a significantly increased melting

temperature with respect to UO2and PuO2

(recom-mendations by Ronchi et al.20,38and Martin et al.25,49)

The standard entropy, based on measurements

of the low-temperature heat capacity, is

S0(298.15)¼ 65.23 J K1mol1.50Bakker et al.20havefitted earlier measurements of H(T)H0(298 K)under the constraint Cp0 (298 K)¼ 61.76 J K1molfrom the low-T heat capacity50to obtain a functionfor the integrated high-temperature heat capacity:

of other actinide oxides, which has been related to apremelting l-transition at 3090 K.38Similar disconti-nuities in the slopes of the enthalpy–temperaturecurves at about 0.8 of the melting temperature werefound for a series of thoria–urania solid solutions.51Recently, Dash et al have performed extensivemeasurements on Th1yUyO2 for y  0.2 (127–

1698 K), making a comparison to older data.52 Theyarrive at the following expression for the heat capac-ity as a function of T (298  T  2000 K) and

y (0.019  y  0.9):

Cp ¼ 66:26 þ 10:91yð Þ þ 0:00923  0:00065yð ÞT

 7:70
105þ 6:7
105y

T2The correlation of Dash et al for pure ThO2 corre-sponds reasonably well to that obtained by Bakker et al.With regard to the mixed oxides, it can be said thatboth (Th,Pu)O2 and (Th,U)O2 experience a slightmelting point depression For (Th,U)O2,the meltingpoints correspond to that of ideal solid solutions;for (Th,Pu)O2,not enough data exist to confirm thisideality.20The Cp(T) data for the (Th,U)O2are quitewell reproduced by a weighted average of the Cp(T)values of ThO2 and UO2, with little deviation.20,52The IAEATechnical Document21uses Bakker’s corre-lation for ThO220and Fink’s correlation for UO239toconstruct Neumann-Kopp heat capacities

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.53This reduces the oxygen

chem-istry of mixed thorium oxides to the chemchem-istry 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 1200C, the maximum

amount of interstitial oxygen increases from 0 for

pure ThO2to 0.25 for Th0.1U0.9O2.25.54This 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 PuO2in the ThO2matrix

simi-larly allows for the creation of oxygen vacancies at

elevated temperatures under inert or reducing

atmo-spheres It may be clear that, because the additive

(UO2or PuO2) will be fissioned away during

irradia-tion 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

Schram53has used the above reasoning to describe

(Th,U)O2+xwith a Lindemer–Bessman type model

for the uranium, which describes the collected

(lim-ited) amount of oxygen potential data reasonably

well On the other hand, Dash et al.52 have used

their heat capacity measurements on Th1yUyO2to

calculate the oxygen potential in a more direct way52

Finally, we note that phase separation occurs foroff-stoichiometric conditions at high U and Pu con-centrations (Section 3.04.4.1) For instance, for high

U concentrations phase separation occurs, and themixed fcc oxide exists in equilibrium with a separate

U3O8phase.54Such a phase equilibrium would duce an oxygen potential plateau

intro-3.04.4 Thorium Oxide Fuel Fabrication

Fabrication of thorium-based oxide fuel is well oped Three routes have been applied successfully tocreate thorium-based oxide fuel: conventional bin-derless powder pressing, spheroidization of powder–plasticizer mixtures, and the sol–gel process Thelatter two cases yield microspheres with diameters

devel-in the range 50–1000 mm, which can be pressed devel-intopellets, used directly in Sphere-Pac/Vipac arrange-ments (see Chapter3.11, Sphere-Pac and VIPACFuel), or coated with carbon and silicon carbidelayers to create TRISO fuel for HTR (seeChapter

the available information is from the Indian ence in powder pressing58 and from the German–American developments in sol–gel methods.2,4,6–10

3.04.4.1.1 ThO2The procedures for fabricating ThO2pellets by pow-der compaction are derived from the proceduresdeveloped for UO2 and (U,Pu)O2fuel.2,10However,

as thorium is found only as a 4-valent cation, it is not

as important to control the oxygen potential duringsintering as in the case of uranium or plutonium, andsintering of thorium oxide may be performed in bothoxidizing and reducing conditions (air, argon, vac-uum, or Ar/H2) On the other hand, the thermal andchemical stability of ThO2discussed in the previous

Figure 3 Oxygen potentials of Th1yU y O 2+x at 1473 K

and different values for x and y, as determined by Schram 53

based on a Lindemer–Bessman model for UO 2+x (black),

and as determined by Dash et al 52 from direct

measurements on the mixed oxides (red).

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

pre-cipitates from a nitrate feed solution (nitric acid with a

pH0.8) upon dropwise addition of oxalic acid:

Th NOð 3Þ4ð Þ þ 2Caq 2O4H2ð Þaq

! Th Cð 2O4Þ2ð Þ þ 4HNOs 3ð Þaq

Early work on powders produced through different

routes (direct ignition of the thorium nitrate,

decom-position of the hydroxide or carbonate) resulted in

large-grained powders that did not sinter well

Calcination is performed in air at 800–900C

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.6,7,10Premilling improves the

sinter-ability considerably, but White et al have observed

that pellets with a density of 96% TD can also be

prepared without premilling when the oxalate

pre-cipitation step is carried out below room temperature

(typically 0–10C).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

(1600C, 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

Follow-ing precompaction at around 100 MPa and

subse-quent granulation, the obtained granules were sieved

through a 14 mesh.6,10

Final compaction of ThO2pellets 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

dif-ferent valency The substitution introduces oxygen

vacancies or interstitials, which enhances the

diffu-sion of thorium ions thereby producing more

homo-geneous and higher-density pellets.7,58Ca2+or Mg2+

(added at1 wt% to the feed solution as a sulfate or

nitrate yielding0.5 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 1700C60 down to 1150–1450C.10 As the

divalent additives introduce oxygen vacancies, they

tend to be better sintering aids under a reducingatmosphere,58while the niobate functions best in anoxidizing atmosphere We refer the reader to Kutty

et al.58for an analysis of the effects of dopants.Finally, it is well known that water easily adsorbs

to the ThO2surface, chemically by forming a highconcentration of hydroxyl groups and physically viahydrogen bonds.61–63Care 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 toreplace standard UO2 or (U,Pu)O2 fuel in LWRs,roughly 25% of LEU or 8–9% LWR-grade Pu or5% WG-Pu or HEU should be added as a ‘topping’ tothe thoria matrix The mixed oxide may be preparedsimply by mixing the separate oxide powders Toenhance homogeneity of the pellets, the thoriumnitrate solution can be mixed with either uranylnitrate or plutonium nitrate before coprecipitation

by the addition of oxalic acid or bubbling of NH3.Before carrying out the coprecipitation step, uranyl-nitrate should first be reduced by the addition ofhydrazine64or by hydrogen gas in the presence of aPt/Al2O3 catalyst For (Th,U)O2 with significantamounts of UO2, the optimum calcination tempera-ture with respect to sintering behavior shifts from 800

to 900C to around 700C.65The fact that uranium and plutonium are multi-valent suggests that the (Th,U)O2 and (Th,Pu)O2mixed oxide are more easily sintered than pureThO2 under oxidative and reductive atmospheres,respectively Indeed, the addition of 2 wt% U3O8 toThO2was found to lower the sintering temperature

in air to 1100C, as was observed for Nb2O5; PuO2yields results similar to Ca2+ in reducingenvironments

InSection 3.04.3.1it has already been mentionedthat the stoichiometric (Th,U)O2 and (Th,Pu)O2forms a single fcc phase in the entire compositionrange, but that in practice off-stoichiometry resultingfrom fabrication conditions results in phase separa-tion at high Pu or U content This should be avoided

as, for instance, the U3O8 phase has a 30% highervolume than the fluorite phase, which results in grainboundary separation and powdering of the fuel.66Kutty et al have studied the sintering behavior ofthe (Th,U)O2.58,64For ThO2–PuO2with Pu contents

up to 30% sintered in Ar or Ar/H2, the sinteredpellets were found to be monophasic, whereas for

Pu contents of 50% and 75%, the ThO–Pu O bcc

phase was found to be present besides the ThO2–

PuO2 fcc structure.58 Similarly, XRD data for both

ThO2–30%UO2and ThO3–50%UO2sintered 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

mate-rials India is developing a method to produce

micro-spheres for HTR TRISO particles (seeChapter3.06,

TRISO Fuel Production) or sphere-pac (seeChapter

3.11, Sphere-Pac and VIPAC Fuel) fuel from ThO2

powder in which a binder material is added In this

so-called CAP (coated agglomerate pelletization) process,

ThO2powder is mixed with a plasticizer (e.g., a

paraf-finum/petrolatum mixture) at slightly elevated

tem-peratures, after which the plastic mixture is simply

molded in an extruder and subsequently in a

spheroi-dizer to form small spheres.67–69In a second step, the

ThO2spheres 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

con-ventional 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

con-trol of the process is rather complex, which may give

problems when scaling up the process

As is the case with the powder compaction

pro-cess, 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

ammo-nia, 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 UO fuel As in the powder–pellet

route, 1 wt% of Ca(NO3)2 may be added to theheavy metal solution as a sintering aid, while 30 gcarbon black per mol heavy metals may be added

to produce spherical pores during the calcinationstep.6–8

Several sol–gel routes have been applied inthe past, of which two have been most successful.The KEMA internal gelation process developed

in the Netherlands for uranium and plutonium hasbeen adapted for thorium in India and Germany,70,71while the external gelation of thorium (EGT) or KFAprocess was developed in Germany (Ju¨lich)6and theUnited States (Oak Ridge)

In the internal gelation process, the nitrate feedsolution is mixed with a solution of hexamethylenetetramine (HMTA, (CH2)6N4) and urea (CO(NH2)2)

of similar concentration at a temperature of around

0C71–73; upon mixing at this low temperature, theheavy metal ions form complexes with urea Theresulting mixture is dispersed as fine droplets by ahollow vibrating needle (frequency in the order

102–103Hz) The dispersed droplets fall into a hot(50–90C) bath of silicone oil, and the droplet temper-ature rises quickly This causes decomposition of theheavy metal–urea complexes as well as hydrolyticdecomposition of HMTA The latter produces ammo-nium hydroxide After hydroxylation of the heavymetal ions by NH4OH, the resulting heavy metalhydroxides form agglomerates of microcrystallites.This induces the sol to gel transition

Details of the reaction for thorium (and also for(Th,U)O2with up to 10% U) have been described byKumar71and Pai et al.74In brief, the concentrated Th(NO3)4solution is partially preneutralized by addingformaldehyde or ammonia The optimum Th4+con-centration after mixing is 1–1.4 M, and the ratio of bothHMTA and urea to Th4+ions is 1.4:1 (Figure 4) Theformation of opaque hard gel spheres upon gelation at

a temperature of 60–70C is taken as an indication thatcrack-free spheres will be formed after drying andcalcination.71–73 Following the gelation step, the gelsare prewashed using CCl4to remove the oil

In the external gelation method, no producing additive (HMTA) is used to hydrolyze orpolymerize the heavy metal sol.70Instead, after beingreleased from the vibrating needle, the sol drops passthrough gaseous ammonia, which quickly gels the sur-face of the droplet The partially gelled drops fall into asolution with a pH of8 containing 1% ammonia and

ammonia-4 M NH4NO3, which completes gelation of the inside

of the droplets As for the internal gelation process, thebest results were obtained when the heavy metal

solution (90C) 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% NH3solution, to improve the internal

structure and to remove organic material, respectively

The particles are then dried at 100–400C in humid

air to avoid crack formation, calcined at around 700C

in air, and finally sintered at 1000–1200C 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 UO2and MOX;Chapter

2.18, Radiation Effects in UO2andChapter2.19,

Fuel Performance of Light Water Reactors

(Ura-nium Oxide and MOX)

3.04.5.1 Neutronic Properties ofThorium-Based Fuel

Although uranium is ‘directly’ fissile, one should bear

in mind that the fissile content (235U) in naturaluranium is only 0.7% In most reactor types, enrich-ment is needed before it can be used to generateenergy 232Th should instead be compared directlywith238U Both are fertile isotopes that are converted

to fissile material in the reactor core In the case of232

Th, the resulting fissile isotope is233U; in238U, it

is 239Pu It is therefore useful to compare the tronic properties of the two sets of nuclides anddiscuss their differences The top and bottom panels

neu-inFigure 5compare the neutron capture cross-sections

of fertile nuclides and the fission cross-sections offissile nuclides, respectively InTable 3, some neu-tronic properties of the relevant nuclides are listed;the table is an abstract from Kaye and Laby.75Regarding the fertile materials, thermal capture isalmost 3 times higher for 232Th than for 238U, butresonance capture is more than 3 times higher for238

U In the fast region of the spectrum, the sections are similar More important for thermalbreeders are the characteristics of the fissile nuclides

cross-It can been seen inFigure 5 that the fission section of233U is least dependent on neutron energy,being relatively small in the thermal region andrelatively large in the epithermal and fast regions.Table 3 shows that the resonance integral for233

cross-U fission is more than 2 times larger than that for239

Pu However, the most significant advantage of233

U compared with235U and239Pu is the very highfission-to-capture ratio (sf/sc), which in a thermalspectrum is about 10 for233U but only about 2.5 for239

Pu.2 This produces a high neutron yield perabsorption  ¼ nsf/(sfþ sc) in a thermal spectrumand up to energies of about 100 keV, that is, it yields arelatively good neutron economy in a wide range ofspectra and especially in thermal and epithermalreactors The value of  also decreases less withtemperature compared to235U and239Pu

On the other hand, a significant drawback of theTh–U cycle is the larger time constant for b-decay ofthe intermediate species: 27 days for233Pa comparedwith 2.3 days for 239Np This has consequences forreactor physics and handling of spent fuel Th-fueledreactors with high neutron densities would containsignificant levels of 233Pa due to its slow decay, andbecause it possesses a relatively high absorptioncross-section, the protactinium will act as a ‘neutronpoison.’ The thorium cycle therefore benefits from

Cracked during processing

Yielded defect-free microspheres

Too soft for processing

Figure 4 Gelation field diagram for ThO 2 as published

by Kumar 68