Analysis of thorium and uranium fuel cycles in an iso-breeder lead fast reactor using extended-EQL3D...

Annals of Nuclear Energy 53 (2013) 492–506

Contents lists available at SciVerse ScienceDirect

Annals of Nuclear Energy

journal homepage: www.elsevier .com/locate /anucene

Analysis of thorium and uranium fuel cycles in an iso-breeder lead fast reactorusing extended-EQL3D procedure

Carlo Fiorina a, Jiri Krepel b, Antonio Cammi a,⇑, Fausto Franceschini c, Konstantin Mikityuk b,Marco Enrico Ricotti a

a Politecnico di Milano, Department of Energy, Nuclear Engineering Division – Via La Masa 34, 2056 Milan, Italyb Paul Scherrer Institut, Nuclear Energy and Safety, Laboratory for Reactor Physics and Systems Behaviour – PSI WEST, Villigen, Switzerlandc Westinghouse Electric Co., Cranberry Township, Pittsburgh, PA, USA

a r t i c l e i n f o

Article history:Received 29 March 2012Received in revised form 28 August 2012Accepted 11 September 2012Available online 5 December 2012

Keywords:Lead fast reactorThorium fuel cycleEQL3DRadio-toxicitySafety parameters

0306-4549/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.anucene.2012.09.004

Abbreviations: BEC, Beginning of Equilibrium CycBOL, Beginning Of Life; BR, Breeding Ratio; EEC, EndEquivalent Full Power Years; ELSY, European Lead SYFuel Assembly; FP, Fission Products; FR, Fast Reactor;Lead-cooled European Advanced Demonstration Reapcm, per cent mille; TD, Theoretical Density; Th–Pu239Pu as main fertile and fissile isotopes; Th–U, core fmain fertile and fissile isotopes; TRUs, TRansUranic iso238U and 239Pu as main fertile and fissile isotopes.⇑ Corresponding author. Tel.: +39 02 2399 6332.

E-mail address: [email protected] (A. Camm

a b s t r a c t

Use of thorium in fast reactors has typically been considered as a secondary option, mainly thanks to apossible self-sustaining thorium cycle already in thermal reactors and due to the limited breeding capa-bilities compared to U–Pu in the fast neutron energy range. In recent years nuclear waste managementhas become more important, and the thorium option has been reconsidered for the claimed potential toburn transuranic waste and the lower build-up of hazardous isotopes in a closed cycle. To ascertain theseclaims and their limitations, the fuel cycle isotopic inventory, and associated waste radio-toxicity anddecay heat, should be quantified and compared to the case of the uranium cycle using realistic core con-figurations, with complete recycle of all the actinides. Since the transition from uranium to thorium fuelcycles will likely involve a transuranic burning phase, this transition and the challenges that the evolvingfuel actinide composition presents, for instance on reactor feedback parameters, should also be analyzed.In the present paper, these issues are investigated based on core physics analysis of the Lead-cooled FastReactor ELSY, performed with the fast reactor ERANOS code and the EQL3D procedure allowing full-corecharacterization of the equilibrium cycle and the transition cycles. In order to compute radio-toxicity anddecay heat, EQL3D has been extended by developing a new module, which has been assessed against ORI-GEN-S and is presented here. The capability of the EQL3D procedure to treat full-core 3D geometriesallowed to explicitly account for aspects related to core dimensions and safety parameters in the analysis,giving a better insight into the pros and cons of the thorium option.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Thorium fuel cycle has been widely studied in the past as a pos-sible alternative to the use of uranium (Kazimi et al., 1999; Kuegleret al., 2007; MacDonald and Lee, 2004; Robertson, 1965), mainlyfor the development of a thermal breeder. Limited studies havebeen carried out concerning the use of thorium in Fast Reactors(FRs) (Tommasi in Gruppelaar and Schapira (2006) and Till et al.

ll rights reserved.

le; BOC, Beginning Of Cycle;of Equilibrium Cycle; EFPY,

stem; EOC, End Of Cycle; FA,HN, Heavy Nuclides; LEADER,ctor; LFR, Lead Fast Reactor;, core featured by 232Th andeatured by 232Th and 233U astopes; U–Pu, core featured by

i).

(1980)), historically conceived as breeder reactors, due to the supe-riority of the uranium/plutonium cycle from this standpoint.

Over the course of the years, waste management has emergedas one of the main problems for public acceptance of nuclear en-ergy, while availability of fissile materials is still of limited concern,especially in western countries. The rationale for FR deploymenthas shifted from breeding to burning TRansUranic isotopes (TRUs)generated by the LWR fleet (NEA, 2006; Salvatores and Palmiotti,2011). The use of thorium as fertile material instead of uraniumtypically increases the TRU burning rate (Sartori et al., 2011). Inaddition, the relatively low mass number of 232Th leads to a char-acteristically low TRU inventory when a Th closed cycle is estab-lished (IAEA, 2002, 2005), with potentially beneficial impacts onthe actinide radio-toxicity and decay heat. Some important safetyfeedbacks of the core (especially the coolant reactivity coefficient)are also improved in a thorium cycle (Pilarski and Lecarpentier,2009; Till et al., 1980).

This study compares the performance of the UO2-based andThO2-based fuel cycles in the iso-breeder (breeding gain equal to

http://dx.doi.org/10.1016/j.anucene.2012.09.004

mailto:[email protected]

http://dx.doi.org/10.1016/j.anucene.2012.09.004

http://www.sciencedirect.com/science/journal/03064549

http://www.elsevier.com/locate/anucene

C. Fiorina et al. / Annals of Nuclear Energy 53 (2013) 492–506 493

zero, B.R. �1.0 � 0.0) ELSY (European Lead-cooled SYstem) core(Alemberti et al., 2011) focusing on the equilibrium core but alsoincluding start-up cores and transition cycles. An initial 238U–Pucore has been selected for the UO2 core while, for the ThO2 core,both 232Th–233U and 232Th–Pu start-up options have been consid-ered. The Pu employed for the start-up cores is assumed to be reac-tor-grade Pu recovered from reprocessing of LWR (Light WaterReactor) used fuel. The core equilibrium state is the main casestudied as it conveys the long-term potential of a given feed option.It represents the asymptotic state reached under the assumption ofconstant fuel management (reloading, cooling, reprocessing, feed)and it is independent of the initial core loading. While true equilib-rium for all the actinide vector requires decades or centuries to bereached, the transition for the main isotopes determining safetyand waste generation is completed in a relatively short time span,generally within a reactor’s operating life. In addition, the morecredible start-up cores and the transition cycles to equilibriumhave also been investigated and will be discussed for each option.

Reactor physics analysis supporting this study has been per-formed with state-of-the-art equilibrium cycle methodologies, i.e.the ERANOS-based EQL3D procedure developed at the ‘‘Paul Scher-rer Institut’’ (Krepel et al., 2009). This procedure enables realisticfull-core 3-D simulations of the cycle-by-cycle behavior of a reac-tor. With respect to lumped approaches frequently adopted in lit-erature (Salvatores et al., 2009; Coates and Parks, 2010), use of fullcore 3-D simulations allows direct evaluation of the core safetyparameters, and a top-level assessment of the safety performancerelated to a fuel cycle selection. While computationally expensive,simulating the cycle-by-cycle operation of a reactor allows to con-tinuously update microscopic and macroscopic cross-sections,accounting for self-shielding effects. Compared to equilibriummethodologies where the equilibrium is found from matrix inver-sion (Salvatores et al., 2009), the cycle-by-cycle iteration processmaintains physical meaningfulness for the transitional steps to-ward equilibrium. Different transition scenarios with various ratesof actinide build-up and ensuing impact on the core safety param-eters can thus be conveniently investigated with EQL3D.

Since the present paper aims at characterizing thorium and ura-nium cycles also from a waste management viewpoint, the EQL3Dprocedure has been extended to provide radio-toxicity and decayheat from the main actinides. This extension is presented here to-gether with the assessment studies performed using the ORIGEN-Scode from Scale 5.1 package (SCALE, 2006). The features of the ex-tended-EQL3D procedure have been employed in the present paperto calculate core configuration, isotopic composition, safety param-eters, radio-toxicity and decay heat for the equilibrium and transi-tion cores of ELSY with a thorium or uranium feed.

This paper is divided into six sections. Section 1 is the introduc-tion. Section 2 presents the methodology, focusing on the exten-sion of EQL3D to calculate core radio-toxicity and decay heat.Sections 3 and 4 present the uranium and thorium cores selectedfor the analyses. Section 5 compares the results obtained for thethorium and uranium fuel cycles. Finally, in Sections 6 and 7, theconclusions of the study are drawn and possible future develop-ments are discussed.

2. Methodology

The present paper compares uranium and thorium cycles in theELSY on the basis of the respective equilibrium cores, and of thetransition toward them. The results presented in this paper havebeen obtained by means of the neutronic code ERANOS 2.2N (Rim-pault et al., 2002). In particular, the EQL3D procedure (Krepel et al.,2009) developed at the Paul Scherrer Institut (Switzerland) hasbeen employed. Starting from an initial fuel composition, EQL3Dsimulates the cycle-by-cycle behavior of a reactor. The main

assumptions are constant imposed reactor power, constant massof actinides in the fabricated fuel and constant fuel management.Under these assumptions, the simulated reactor always reachesits final equilibrium state. The resulting reactivity indicates thecapability of the reactor to support a closed fuel cycle: breeder oriso-breeder reactors are expected to show a positive reactivity atequilibrium. The core is represented in its full dimensionality, thusallowing a meaningful characterization in terms of core perfor-mance as well as safety-related parameters, both at equilibriumand during the transition toward it.

The assumption of fixed mass of actinides implies that the den-sity of the remanufactured fuel, from BOL (Beginning Of Life) toequilibrium, remains constant. In principle, the fuel theoreticaldensities will vary through the recycles due to the changing fuelactinide composition. As far as oxide fuels are concerned, pureThO2 and PuO2 have theoretical densities equal to 10.0 g/cm3 and11.46 g/cm3, respectively (Orlov et al., 2001; Rodriguez and Sunda-ram, 1981). In addition, the determination of the smeared density(i.e., the fuel mass divided by the volume inside the active part ofthe pins) depends on the fuel form and manufacturing technique,which is still speculative at this stage. Also the smeared densitywill likely change through the cycles of manufacturing as a resultof challenging conditions from increasingly radioactive fuel andHe release from alpha decay of higher actinides. Smeared densitiesare typically 80–90% of the Theoretical Density (TD). For simplicity,a smeared density of 87% TD (Alemberti et al., 2011), and the UO2

theoretical density of 10.96 g/cm3 (Orlov et al., 2001) were as-sumed for all cases considered. Uranium is the main componentin a U–Pu core and the second in a Th–U core. The value adoptedis between ThO2 and PuO2 densities, thus representing a reason-able value for the Th–Pu core. While the adopted value likelyunderestimates the U–Pu fuel density and overestimates that ofTh–U, the impact on the core actinide content is expected to bemarginal. In fact, in Fiorina et al. (2011), the actinide content nec-essary for a thorium iso-breeder was found to be fairly indepen-dent of the actinide density of the adopted fuel. The assumedfuel density will impact core dimensions, leading to a slightly lar-ger difference between the iso-breeder U–Pu and Th–U core heightthan that calculated in this paper.

Full recycle of actinides has been assumed for this study, i.e.during reprocessing fission products are removed, all actinidesare recycled and either natural thorium or uranium is added asfeed until the actinide mass of the initial fresh fuel is restored.Due to the scoping nature of the calculations performed, a single-batch irradiation scheme has been assumed for convenience. Un-der this assumption, the entire core is irradiated for the full fuelirradiation time, unloaded (and ideally replaced with anotherone), cooled for an equally long period, reprocessed, and reloadedonce again. For simplicity, the irradiation time will then be referredto as cycle. The approximations related to the use of a single-batchirradiation scheme are discussed in some details e.g. by Artioliet al. (2009) and Krepel et al. (2010).

Full core flux and burn-up calculations have been performedwith ERANOS in the 33 energy-group structure optimized for FRcalculation. The multigroup nodal transport theory code VARIANThas been used for flux calculations (Ruggeri, 1999), employing aP3 approximation with simplified spherical harmonics. The 33-group cross-sections have been obtained from assembly-wise lat-tice calculations using the collision-probability code ECCO in1968 energy groups based on the JEFF3.1 library available in ERA-NOS (Sublet et al., 2006). The ECCO lattice calculations have beenperformed with the consistent solution method (Rimpault, 1997).

Each Fuel Assembly (FA) has been discretized in 8 axial nodesfor fuel depletion calculations. Evolution of masses is computedfor each of the 8 nodes of each FA according to the specific powerderived through the full core flux calculations. During each cycle,

Table 1Isotopes selected for radio-toxicity and decay heat calculations for the lower part ofthe decay chains.

Chain Selectedparentisotope

Parentisotopehalf-life

Longest-livedisotope in theprogeny

Half-life of thelongest-livedisotope in theprogeny

Radium 226Ra 1600 years 210Pb 22.3 yearsActinium 227Ac 21.8 years 227Th 18.7 daysThorium 228Th 1.9 years 224Ra 3.7 daysNeptunium 229Th 7340 years 225Ra 14.9 days

494 C. Fiorina et al. / Annals of Nuclear Energy 53 (2013) 492–506

fluxes are recalculated 9 times. Between two flux recalculationsmacroscopic cross-sections are computed 9 times, and the specificpower of each node is accordingly renormalized to maintain a con-stant core power. The microscopic cross-sections are calculatedevery few cycles, the exact number depending on the rate of vari-ation of the fuel composition. Three sets of microscopic cross-sec-tions are calculated, one for each fuel zone (inner, middle, outerzone – see Section 3).

2.1. Isotope selection for thorium cycle analysis

Consideration has been given to selecting the isotopes whoseevolution had to be explicitly simulated. Calculations have beeninitially performed considering as many as 56 heavy isotopes,and 179 Fission Products (FP). In particular, all the actinides lighterthan 252Cf and available in ERANOS JEFF 3.1 libraries were used, ex-cept for 235Np, 249Cm, 250Cm, 247Bk, 250Bk.

Due to the scoping nature of the present calculations, and to re-duce the computational burden, it was eventually decided not toconsider explicit FP, but to limit the analysis adopting the 18 ‘‘glo-bal pseudo fission products’’ prepared for Sodium Fast Reactorsand available in the ERANOS libraries. Such fictitious isotopes werecreated to account for the main reactivity effects due to FP in a FR(Tommasi, 2001). Use of pseudo-fission products was found to af-fect mildly the obtained results (e.g., an underestimation in the300–500 pcm range was found for a typical equilibrium reactivityin the ELSY, for uranium and thorium cycle, respectively). On theother hand, this choice allowed a reduction of the computationaltime by a factor of three. A further assessment of the impact ofpseudo-fission products on the EQL3D procedure calculations isavailable in (Krepel et al., 2009). For what concerns radio-toxicityand decay heat, FP are known to play a secondary role in the longterm (Salvatores and Palmiotti, 2011).

The selection of the heavy nuclides was performed based ontheir impact on reactivity, radio-toxicity and decay-heat. As a re-sult, it was decided to consider the following 41 isotopes: 226Ra;227Ac; Th from 228 to 233, excluding 231Th; Pa from 231 to 233;U from 232 to 238; Np from 237 to 239; Pu from 238 to 242;Am from 241 to 243; Cm from 242 to 248; 249Bk; and Cf from249 to 252. With respect to other actinide chains commonlyadopted in literature for the analysis of the uranium cycle (e.g.,by Krepel et al. (2009)), 226Ra, 227Ac, 228Th and 229Th were includedto calculate the evolution of radio-toxicity and decay heat (see nextSection). 233Th has major effects on in-core decay heat (see Sec-tion 5.2.2). 232Pa is parent of 232U and was considered due to itsimportance in reprocessing and non-proliferation issues. 237U rep-resents one of the gateways for the production of TRUs in the tho-rium fuel cycle. Finally, 230Th was considered because it is presentin non-negligible amounts in an equilibrium thorium cycle (see Ta-ble 4). The resulting chain is very similar to the one proposed inCoates and Parks (2010). In particular, with respect to Coates andParks, the actinide chain is extended above 244Cm and below230Th, but it bypasses 231Th, 231U, 239U, 243Pu, 244Am. The isotopes231Th, 231U and 239U are not available in the ERANOS JEFF 3.1 li-brary, while 243Pu and 244Am are present in negligible amountsin the core (in the order of few mg in a 50 t inventory) even in afully closed uranium equilibrium cycle. As concerns reactivity atequilibrium, use of the 41 heavy nuclides (instead of 56) led toan overestimation in the order of few pcm in the equilibrium core(uranium or thorium cycle), but allowed a further 10–15% reduc-tion of the computational time.

2.2. Extended-EQL3D: radio-toxicity and decay heat calculations

The original EQL3D procedure did not compute radio-toxicityand decay heat of the fuel. For this reason, it was decided to modify

it and develop a dedicated module. This new module uses the iso-topic composition from EQL3D to perform a simplified decay, keep-ing track of the more relevant isotopes, and applying the adultingestion coefficients from the ICRP72 (ICRP, 1996), to calculateradio-toxicity. The decay heat is computed using the Q-value ofthe reactions. This is correct for alpha decay and represents a goodapproximation for gamma decay. The use of Q-value for beta decayleads to a systematic overestimation because the portion of the en-ergy carried away by neutrinos is not subtracted. Beta decay is par-ticularly important during the reactor operation, but it becomesunimportant after few months of cooling (Salvatores, 2002).Accordingly, decay heat calculations in the present paper are ex-pected to be correct for long term cooling, but will overestimatethe in-core decay heat. This is consistent with the choice of model-ing only actinides, without explicit fission products, focusing theattention on the middle and long term evolution of the spent fuel.

The main obstacle which was encountered while setting-up thenew module is related to the limited number of isotopes availablein the ERANOS JEFF3.1 library. This library was created for reactorcalculations and only isotopes heavier than radium are availablewhile, for radio-toxicity calculations, all the isotopes in the decaychains down to the stable isotopes need to be considered. Thereare four main actinide decay chains: the radium, actinium, thoriumand neptunium chains, which end in the stable isotopes 206Pb,207Pb, 208Pb, 209Bi, respectively (Krane, 1988). The decay time ofthe isotopes in the lower part of these chains (i.e., close to the sta-ble isotopes) generally decreases with decreasing mass number, sothat for each chain it was possible to select one parent isotopewhich: (1) is available in the adopted libraries, and (2) whosehalf-life is orders of magnitude longer than those of the progeny.Table 1 reports the half-lives of these four isotopes, together withthe longest-lived isotope in their progeny.

When the half-life of the progeny is orders of magnitude belowthat of the parent, and after a period of time equal to 6–7 times thehalf-life of the longest-lived isotope in the progeny, parent andprogeny approximately reach a ‘‘secular equilibrium’’ state (Krane,1988) where each isotope has the same activity. In the present pa-per, the 4 parents listed in Table 1 are considered to be always insecular equilibrium with the progeny. The activity of the parents,that are all available in the JEFF3.1 ERANOS library, is then usedfor computing radio-toxicity and decay heat of the progeny, thusovercoming the incompleteness of the library. Secular equilibriumis also assumed for the other isotopes not directly simulated. Inparticular, 228Ra and 228Ac are considered to have the same activityas 232Th, and 234Th and 234Pa the same as 238U. 241Cm, 231Th, 243Puand 244Pu are instead considered to have the same activity as241Am, 235U, 247Cm and 248Cm, respectively.

As mentioned, secular equilibrium is approximately reachedafter a period equal to 6–7 times the half-life of the longest-livedisotope. This means that after �100 days it is possible to correctlypredict radio-toxicity and decay-heat of the progeny of 227Ac, 228Thand 229Th by making use of the secular equilibrium assumption.Approximately 100 years are instead necessary for considering


226Ra in equilibrium with the progeny. Assumption of secular equi-librium for shorter periods of time is then expected to lead to anoverestimation of the activities and, consequently, of radio-toxicityand decay heat. In addition, the condition of secular equilibrium isapplicable if an initial given amount of the parent nuclide is con-sidered. In the present case, the chosen 4 parents are continuouslyfed by the decay of heavier nuclides. This leads to a systematicoverestimation of the progeny activities when calculated assumingsecular equilibrium. Related inaccuracies are expected to increasewith the feed rate of each parent. Clearly, the adequacy of the sec-ular equilibrium assumption is strongly dependent upon the par-ticular situation considered (relative importance of the isotopesinvolved in the approximation, rate of mass variation, etc.). Itsvalidity for the sets of simulation performed will be assessed inSection 5.2, by benchmarking its predictions against the results ob-tained with ORIGEN-S.

Fig. 1. ELSY core (a) and assembly (b) designs (Alemberti et al., 2011).

3. U–Pu ELSY core

The ELSY fast reactor concept (Alemberti et al., 2011) is adoptedas reference core design in the present paper. It was designed dur-ing the EURATOM sixth framework program (2002–2006) and fur-ther developed in the LEADER (Lead-cooled European AdvancedDemonstration Reactor) project of the seventh framework program(2007–2011). It is an iso-breeder (breeding gain equal to zero) LeadFast Reactor (LFR) for electricity production and capable of minoractinide burning. It was designed to work in a classical U–Pu fuelcycle. The ELSY design presented in (Alemberti et al., 2011) hasbeen chosen for this paper. Its core is composed by hexagonal fuelbundles with surrounding duct (wrapper). Characterizing parame-ters are summarized in Table 2, while a schematic representationof core and FA designs is reported in Fig. 1. In the following, thiscore design used in a U–Pu cycle will be referred to as ‘‘U–Pu’’.

The EQL3D procedure has been applied to the ELSY core in orderto simulate its evolution in a closed fuel cycle. The initial composi-tion was assigned using a ratio of plutonium to natural uraniumclose to the equilibrium one and leading to a reactivity roughlyequal to 1 (keff � 1.000) at the End Of Cycle (EOC). The plutoniumisotopic composition adopted is from Artioli et al. (2007). Starting

Table 2Main core parameters of the ELSY nominal core.

Unit Value

Electric power MWe 600Thermal power MWth 1500Number of assemblies for inner/middle/

outer zone– 163/102/168

Effective core radius for inner/middle/outer zone

m 1.42/1.81/2.31

Fuel active height m 1.2Average lead temperature K 713Average fuel temperature K 1223Fuel type – U–Pu oxideFuel smeared density %TD 87Fuel theoretical density g/cm3 10.96Plutonium isotopic composition

238Pu/239Pu/240Pu/241Pu/242Puwt% 2.332/56.873/26.997/

6.104/7.693Peak-to-average FA power ratios at BOL – 1.18Position of the peak power FA at BOL – First FA row of the

outer fuelPeak-to-average FA power ratios at EOC – 1.52Position of the peak power FA at EOC – Core centerCladding material – Steel T91Fuel residence time/fuel cycle EFPY* 6Refueling time days 30

* Equivalent Full Power Years.

from the initial core, one hundred cycles (600 years irradiationtime) were simulated to assure full convergence to equilibrium.Fig. 2 shows the mass evolution of some important isotopes forthe first simulated 150 EFPY and their different rate of convergenceto the respective equilibrium concentration. The jigsaw behaviorobserved for some of the isotopes is due to their decay in-betweencycles and subsequent build-up during irradiation. It is interestingto observe that at the end of the 150-year irradiation period theplutonium composition has reached the equilibrium values. Hea-vier actinides still experience a variation but are virtually at equi-librium. 252Cf, the heaviest isotope considered and expected toconverge last, has reached 90% of its equilibrium mass at the endof the 150 EFPY irradiation time.

Fig. 3 shows the respective evolution of the core multiplicationfactor (keff). The typical breeding behavior is noticeable. A slightlypositive equilibrium reactivity is established for the UO2 ELSY corein line with the iso-breeder design. The transition toward the equi-librium reactivity is completed in a typical reactor lifetime (50–60 years), consistently with the evolution of the main isotopes(Fig. 2).

Fig. 4 plots the assembly-wise power distribution in the core atBOL, EOC and EEC (End of Equilibrium Cycle). The initial fissile con-tent in the different core regions has been purposely set-up to re-duce power peaking but the power distribution worsens duringirradiation, a trend that can be counterbalanced by adopting a mul-ti-batch scheme (Artioli et al., 2009).

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

0 50 100 150

Mass of C

f252 [kg]

Mas

s [k

g]

Pu238

Pu239

Pu240

Pu241

Pu242

Am241

Cm244

Cf252

0.0E+00

2.0E+02

4.0E+02

6.0E+02

8.0E+02

1.0E+03

1.2E+03

0.E+00

1.E+03

2.E+03

3.E+03

4.E+03

5.E+03

6.E+03

0 50 100 150

Other m

asses [kg]

Mas

ses

of P

u239

and

Pu2

40 [k

g]

Time [EFPY]

Pu239

Pu240

Pu238

Pu241

Pu242

Am241

Time [EFPY]

Cm244

(a)

(b)

Fig. 2. Mass evolution over 150 EFPY of U–Pu closed cycle in the ELSY core with: (a)logarithmic and (b) linear scale for masses.

0.99

1

1.01

1.02

1.03

1.04

1.05

1.06

1.07

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Keff

[-]

Time [EFPY]

Fig. 3. keff Evolution over 150 EFPY of U–Pu closed cycle in the ELSY core.

0

1

2

3

4

5

6

0 50 100 150 200 250

Asse

mbl

y po

wer

[MW

]

Assembly radial position [cm]

U-Pu BOL

U-Pu EOC

U-Pu EEC

Fig. 4. Assembly-wise radial power distribution for U–Pu closed cycle in the ELSYcore.


4. Proposed iso-breeding thorium ELSY core

This Section describes the modifications made to the ELSY coreto maintain its iso-breeder behavior in a thorium cycle. In this way,the new core can be operated with 232Th as the only feed and main-tain positive reactivity at equilibrium. In line with the nominal U–Pu core, thorium oxides (Aronson et al., 2006; Benedict et al., 1981;Rodriguez and Sundaram, 1981) were selected as fuel form. Samepin and assembly designs as for the U–Pu counterpart have beenadopted for the ThO2 core (see Fig. 1b).

Replacing ThO2 feed with UO2 feed without any core modifica-tions leads to a strongly negative equilibrium reactivity (Fiorinaet al., 2011), as a consequence of the inferior breeding perfor-mance. In order to maintain an iso-breeding configuration, the coredesign had to be modified. For simplicity, it was decided to in-

crease the core height until the criticality (keff � 1.000) at EECwas reached. Increasing the core height reduces axial neutron leak-age, increases the fuel inventory and reduces the specific power.After some iteration, it was found that the core height needs tobe increased from 1.2 m to 1.7 m. The choice of increasing the coreheight is not necessarily optimal from a design viewpoint. It min-imizes leakages compared to other options, which in turn mini-mizes the actinide inventory necessary to achieve iso-breeding.Increasing the core height also allows to maintain lead velocitiesand lead axial temperature increase in the core, two of the mainconstraints in the ELSY design. This choice is convenient to illus-trate the various effects at play without proceeding to a full coreredesign, which is out of the scope of the present work. The effectof the core geometry on the presented results will be briefly dis-cussed in Section 5.1. The modified thorium core will be fromnow on identified as ‘‘Th–U’’ when used in a ‘‘pure’’ (i.e., with onlyTh feed and 233U as initial fissile loading) thorium cycle.

The discharge burnup in the ELSY is limited by fuel claddingdamage (<100 dpa (Sobolev et al., 2009)). The standard U–Pu corecomplies with this constraint. In order to maintain the claddingdamage within acceptable limits, the average discharge burn-up(64 GWd/tHM) of the U–Pu core has been adopted for the Th–Ucore. This choice led to �8.5-year fuel irradiation time, comparedto the 6 years of the U–Pu option. Given the reduced specific powerof the thorium core design due to the taller core, the longer irradi-ation time still leads to a cladding fluence �10% below that of theU–Pu fuel. In view of the similar spectrum, cladding damage is alsoexpected to remain smaller than, or comparable to, the U–Pu case.

An important issue is represented by the transition scenariofrom uranium to thorium cycle. 233U is not available in natureand would require dedicated irradiation of 232Th fuel, presumablyin blankets of U–Pu cores, for its generation. The implementationof a thorium cycle with 233U as fissile material would then be pos-sible only after irradiation of 232Th and recovery of 233U, with longwaiting times and possible proliferation concerns. A more likelysecond option would be to start using thorium with a different fis-sile material, with in-bred 233U gradually replacing it in the subse-quent cycles. Use of enriched uranium is a possibility. If denaturedwith 238U, this choice would have the advantage of initiallyincreasing proliferation resistance by denaturing the in-bred 233Ubut would worsen the fuel cycle performance and the transition to-ward equilibrium through the production of Pu from 238U. A morelogical choice in terms of waste minimization would be to initiallyload the reactor with recycled Pu from LWRs which would contrib-ute to eliminating some of the Pu legacy. This initial Th–Pu core,

1.E-10

1.E-09

1.E-08

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

0 50 100 150

Mass of C

f252 [kg]

Mas

s [k

g]

Time [EFPY]

Pu238 (Th-U)

Pu238 (Th-Pu)

U234 (Th-U)

U234 (Th-Pu)

U233 (Th-U)

U233 (Th-Pu)

U232 (Th-U)

U232 (Th-Pu)

Cf252 (Th-U)

Cf252 (Th-Pu)

0.0E+00

5.0E+01

1.0E+02

1.5E+02

2.0E+02

2.5E+02

3.0E+02

3.5E+02

4.0E+02

0.E+00

1.E+03

2.E+03

3.E+03

4.E+03

5.E+03

6.E+03

7.E+03

8.E+03

0 50 100 150

Mass of U

232 and Pu238 [kg]Mas

s of

U23

3 an

d U

234

[kg]

Time [EFPY]

U234 (Th-U)

U234 (Th-Pu)

U233 (Th-U)

U233 (Th-Pu)

Pu238 (Th-U)

Pu238 (Th-Pu)

U232 (Th-U)

U232 (Th-Pu)

(a)

(b)

Fig. 5. Mass evolution over 150 EFPY of Th–U and Th–Pu cases with a) logarithmicand b) linear scale for masses.

0

1

2

3

4

5

6

7

0 50 100 150 200 250

Asse

mbl

y po

wer

[MW

]

Assembly radial position [cm]

Th-U BOLTh-U EOCTh-U EECTh-Pu BOLTh-Pu EOC

Fig. 7. Assembly-wise radial power distribution for Th–U and Th–Pu cases.


identified in the following as ‘‘Th–Pu’’, will evolve towards the Th–U equilibrium core, assuming a fixed reloading scheme and tho-rium as the makeup feed.

Similarly to the U–Pu case, 100 cycles have been simulatedusing EQL3D. Fig. 5a, b and Fig. 6 show the resulting evolution ofmasses and reactivities for both Th–U and Th–Pu core, for the first150 EFPY. As expected, in both cases the reactor tends to convergetoward the same Th–U equilibrium state. Fig. 5a and b shows thatafter 150 EFPY, the concentration of the main isotopes (232Th, 233U,234U) has converged to equilibrium and, consequently, the reactiv-ity profile has also converged (Fig. 6). The main isotopes impactingreactivity reach the equilibrium value in 3–4 cycles for the Th–Ucase. Unlike in the previous U–Pu case, the Pu and higher actinidesare still not converged. For instance, 238Pu is slowly burned in Th–Pu, while it slowly builds up in Th–U; as a result the 238Pu contentfor the two start-up cores is still different after 150 EFPY. The dif-

0.99

1

1.01

1.02

1.03

1.04

1.05

1.06

1.07

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Keff

[-]

Time [EFPY]

Th-U

Th-Pu

Fig. 6. keff Evolution over 150 EFPY of Th–Pu and Th–U cases.

ference in the convergence becomes larger for the higher massnumber isotopes, underlining their delayed build-up in the Th–Ucore. The 252Cf content in the Th–Pu start-up core is three ordersof magnitudes larger than in the Th–U case after 150 EFPY.

The initial fissile enrichment in the three core zones (Table 2)has been chosen to minimize the power peaking. Fig. 7 plots theresulting assembly-wise power distribution at BOL, EOC and EEC.A peak-to-average FA power ratio equal to 1.2 and 1.15 has beenachieved at BOL for Th–U and Th–Pu, respectively. Such peakingfactors reach at EOC a value equal to 1.39 and 1.65, respectively,which become 1.39 for both at EEC (the equilibrium state is thesame for the two cores). Peak-to-average FA power ratios are sim-ilar to those of the nominal U–Pu core (Table 2) and their notice-able variations are partly related to the single-batchapproximation adopted. A more accurate representation of thereactor by adopting a multi-batch simulation would change thepower distribution but this is expected to have only a minor im-pact on reactivity and fuel inventory evolution. A possible specificconcern of the Th cores is related to the relatively long (27-day)half-life of 233Pa, the precursor of 233U, whose impact on reactivityand breeding gain could be affected by the power distribution. Thispotential impact has been ruled out by performing a second calcu-lation at reduced reactor power but longer irradiation time (thuspreserving fuel burnup) leading to a negligible change in reactivitywith respect to the nominal case.

Finally, it is interesting to observe that the reactivity is alwayspositive in the Th–Pu case, thus indicating that no top-up of fissilematerial would be needed for the transition from cycle 1 to equi-librium. During the first cycle, the reactor increases its reactivityby as much as �4000 pcm. This is a notable reactivity variationto be balanced with control rods, but it can be decreased movingto a 3 or 5 refueling batches. Insertion of dummy elements orThO2 control/absorber rods in the core could also be an option.The initial reactivity increase of the Th–Pu core is attributed tothe larger neutron yield per fission (m) of Pu vs. U, �17% higherfor ELSY, and the ensuing larger initial breeding.

5. Comparison between uranium and thorium cycles

For a meaningful comparison of the thorium and uranium fuelcycles while keeping the problem tractable, a set of three corestates has been selected. The U–Pu equilibrium core is sufficientto characterize the uranium cycle: it represents the natural stateof the reactor, it is relatively close to the start-up core and it isthe limiting case for radio-toxicity and decay heat calculations. Incase of the thorium cycle, both the Th–U equilibrium and theTh–Pu start-up cores are analyzed. The Th–Pu is the likely choice

Table 3Selected core states for the comparison.

U–PuEEC

Th–UEEC

Th–Pu BOL

Core height (cm) 120 170 170Actinide mass (t) 51 72 72Fuel residence time/fuel

cycle (EFPY)6 8.5 8.5

Core state EEC EEC BOLPeak-to-average FA power

ratios (�)1.59 1.39 1.15

Position of the peak powerFA

Corecenter

Corecenter

First FA row of theouter fuel


for the start-up core, has markedly different features than the Th–U equilibrium state, and has thus been included in the comparison.

The rest of the paper will focus on these three cores, includingthe transition from Th–Pu to Th–U. The main features of the se-lected three core states are summarized in Table 3. For the equilib-rium cores, the EEC state is chosen in order to include in theanalysis the effect of fission products. The BOL state is instead con-sidered for Th–Pu in order to exclude from the analysis the effect of233U build-up, which could mask some differences with respect tothe Th–U equilibrium. Table 4 shows the actinide masses for thethree different options. The masses for U–Pu EEC and Th–U EEChave been obtained after the 100-cycle simulation presented inthe previous two Sections. As expected, the main isotopes for theU–Pu equilibrium are 238U and the plutonium isotopes, 238Pu to242Pu. For the Th–U equilibrium, 232Th and the uranium isotopesfrom 233U to 236U are instead predominant. It should be remarkedthe small build-up of transuranic elements in Th–U EEC. Only237Np and 238Pu show amounts which are roughly comparable totheir counterpart amounts in U–Pu EEC. All the other TRUs arethree orders of magnitude less abundant compared to their inven-tory in the U–Pu equilibrium core. On the other hand, a notablebuild-up of 232U can be observed for Th–U EEC. The 232U progenyis characterized by high energy gamma emitters, which greatlycomplicates handling, transporting and manufacturing the repro-cessed fuel but reduces the attractiveness of the in-bred U for pro-liferation intents. In the present case, the 232U equilibriumconcentration in uranium is on the order of 1500 ppm. The conse-quences on radio-toxicity and decay heat will be discussed in some

Table 4Fuel composition (masses expressed in kg).

U–Pu EEC Th–U EEC Th–Pu BOL

226Ra 1.79 � 10�4 1.61 � 10�2 0227Ac 3.38 � 10�5 1.20 � 10�2 0228Th 1.96 � 10�4 3.98 � 10�1 0229Th 8.38 � 10�5 1.14 0230Th 1.18 � 10�10 8.97 0232Th 5.60 � 10�4 5.68 � 104 5.84 � 104

233Th 4.42 � 10�10 3.29 � 10�2 0231Pa 8.82 � 10�3 1.68 � 101 0232Pa 3.95 � 10�6 7.61 � 10�3 0233Pa 2.96 � 10�6 5.71 � 101 0232U 7.64 � 10�3 1.55 � 101 0233U 1.01 � 10�2 6.84 � 103 0234U 1.52 � 102 2.27 � 103 0235U 6.62 � 101 4.96 � 102 0236U 1.24 � 102 5.55 � 102 0237U 9.20 � 10�2 1.70 � 10�1 0238U 3.74 � 104 2.68 � 10�1 0237Np 6.46 � 101 1.24 � 102 0238Np 3.30 � 10�2 4.48 � 10�2 0239Np 4.42 3.07 � 10�5 0238Pu 2.62 � 102 8.50 � 101 3.26 � 102

detail in Section 5.2. From the proliferation viewpoint, the gammadose rate at 0.5 m (a typical working distance for glove-box oper-ations) from a 5-kg sphere of 233U containing 1500 ppm of 232Uis �0.2 Sv/hr (Kang and von Hippel, 2001). Exposed workers wouldreach the annual dose limit in few minutes, and a few hours ofexposure could be life-threatening. In addition, shielding high en-ergy (2.6 MeV) gammas is difficult and the characteristic energeticpeak would make the U easily detectable. However, 24000 ppm of232U in U would be necessary to meet IAEA’s standard for reducedphysical protection requirements (>1 Sv/h at 1 m (IAEA, 1999)).

5.1. Safety and control

A primary concern when considering the adoption of a differentfuel in a reactor derives from its impact on the core safety features,which may change from the initial load through the transition tothe equilibrium cycle. It is practical to focus this preliminary anal-ysis on some lumped parameters representative of the reactorbehavior during both operational and accidental transients. In thissubsection, the U–Pu EEC, Th–U EEC and Th–Pu BOL cores are com-pared employing the core safety parameters summarized in Ta-ble 5: fuel Doppler coefficient, fuel expansion coefficient, diagridexpansion coefficient, coolant expansion coefficients, neutron gen-eration time and beta (core-average delayed neutron fraction)effective. All the coefficients have been calculated through thereactivity variation following a 100 K temperature increase (forthe fuel, the coolant or the diagrid) starting from the nominal oper-ating temperatures in Table 2. Doppler and fuel expansion havealso been condensed in a ‘‘fuel coefficient’’. This is justified bythe fact that both coefficients depend on the same temperature.Void reactivity has been calculated through complete voiding ofthe active core, while coolant expansion coefficient resulted froma uniform expansion of the coolant in the entire core. Diagridexpansion represents a radial core expansion driven by diagrid,which is the steel structure supporting the assemblies in the core.The fuel linear expansion coefficient depends on the fuel type, butdifferences are generally small for the fuels considered (Aronsonet al., 2006; Orlov et al., 2001; Rodriguez and Sundaram, 1981).Therefore, the same linear expansion coefficient, equal to1.2 � 10�5 K�1, has been considered for all three fuels. This choicehelps comparing the neutronic performance of the fuels only,

U–Pu EEC Th–U EEC Th–Pu BOL

239Pu 5.02 � 103 1.94 � 101 7.96 � 103

240Pu 3.42 � 103 7.21 3.78 � 103

241Pu 3.51 � 102 7.29 � 10�1 8.55 � 102

242Pu 3.47 � 102 5.54 � 10�1 1.08 � 103

241Am 3.29 � 102 6.65 � 10�1 0242Am-g 5.15 � 10�2 9.08 � 10�5 0242Am-m 2.52 � 101 4.43 � 10�2 0243Am 1.09 � 102 1.47 � 10�1 0242Cm 1.03 � 101 1.83 � 10�2 0243Cm 1.22 2.01 � 10�3 0244Cm 6.24 � 101 7.47 � 10�2 0245Cm 1.62 � 101 1.65 � 10�2 0246Cm 1.10 � 101 7.91 � 10�3 0247Cm 2.20 1.34 � 10�3 0248Cm 1.63 6.41 � 10�4 0249Bk 2.75 � 10�2 8.10 � 10�6 0249Cf 1.29 � 10�1 4.28 � 10�5 0250Cf 3.91 � 10�2 1.04 � 10�5 0251Cf 9.65 � 10�3 1.98 � 10�6 0252Cf 6.02 � 10�4 1.22 � 10�7 0TOTAL 5.11 � 104 7.22 � 104 7.25 � 104

Table 5Comparison among safety parameters.

U–PuEEC

Th–UEEC

Th–PuBOL

Doppler coefficient (pcm/K) �0.55 �0.79 �0.64Doppler coefficient (10�3 $/K) �1.7 �2.4 �2.3Fuel expansion coefficient (pcm/K) �0.18 �0.13 �0.17Fuel expansion coefficient (10�3 $/K) �0.56 �0.40 �0.62Fuel coefficient (Doppler + fuel expansion)

(pcm/K)�0.73 �0.92 �0.81

Fuel coefficient (Doppler + fuel expansion)(10�3 $/K)

�2.3 �2.8 �3.0

Diagrid expansion coefficient (pcm/K) �0.61 �0.44 �0.40Diagrid expansion coefficient (10�3 $/K) �1.9 �1.3 �1.5Coolant expansion coefficient (pcm/K) 0.53 0.26 0.51Coolant expansion coefficient (10�3 $/K) 1.7 0.80 1.9Void reactivity (pcm) 5540 2415 4801Void reactivity ($) 17.3 7.41 17.6Generation time (ls) 0.77 1.03 0.87Beta effective (pcm) 320 326 273


without the impact of slightly different (and uncertain) expansioncoefficients.

The feedback coefficients listed in Table 5 give important indi-cations about the core behavior, both in operational and in acci-dental conditions. The fuel coefficient is the only ‘‘instantaneous’’feedback acting during power excursions. The coolant expansionshows positive values and represents one of the known concernsof FRs. Generally it has a smaller magnitude than the fuel coeffi-cient and, in case of power excursions, its effects is delayed. Theconcerns about the positive coolant expansion coefficient arisemainly from: (1) the reactor manoeuvrability, due to the negativeimpact on load-following capabilities and, more relevant to safety,(2) the possibility of coolant overheating. In fact, the coolant wouldexpand and cause a positive reactivity insertion and a possibletransient overpower. This could damage the core before the inter-vention of the fuel coefficient and/or of the control system could beeffective. Th–U EEC, featuring notable improvements in both fueland coolant coefficients, demonstrates the potential for enhancedinherent safety characteristics, as well as a better controllability,compared to the counterpart U–Pu EEC. On the other hand, diagridexpansion coefficient is noticeably worsened. This could have anegative impact in accidental transients driven by the inlet coretemperature, e.g., in case of a loss of heat sink. Load-followingoperation could also be disturbed, but diagrid usually features anon-negligible heat capacity and its effect is delayed comparedto coolant or fuel.

The neutron generation time and void reactivity do not affectnormal reactor operations, but may represent a concern in caseof severe accidents. In particular, the reactor period becomes di-rectly proportional to the generation time for reactivity insertiongreater than one dollar. Void reactivity represents a fundamentalparameter mainly in sodium fast reactors, due to the risk of coolantboiling. Nonetheless, also for LFR a partial core voiding can be pos-sible in case of simultaneous and sudden failure of a large numberof fuel pins, with consequent release of helium and fission gases. Asshown in Table 5, both generation time and void reactivity are im-proved in Th–U EEC, which makes such core safer during severeaccidents.

Th–Pu BOL generally lies between the U–Pu and Th–U equilib-rium parameters and has accordingly intermediate advantages,or disadvantages when diagrid expansion plays a role, comparedto U–Pu. An important exception is the beta effective, for whichTh–Pu BOL features a particularly low value. Th–Pu BOL will bequicker in reaching a condition of prompt criticality and, in gen-eral, it will react more rapidly to reactivity insertions. In particular,the Th–Pu coolant expansion coefficient and void reactivity be-come worse than in the U–Pu EEC case when expressed in dollars

(Table 5). On the other hand, generation time and fuel coefficientare improved with respect to U–Pu EEC. The improved fuel coeffi-cient suggests a core behavior which is not worse (if not improved)than the U–Pu core. However transient simulations, which are be-yond the scope of the present work, would be required for a morereliable comparison.

In summary, safety parameters are generally improved in theTh–U EEC state, and comparable for U–Pu EEC and Th–Pu BOL. Inthis sense, it is worth noticing that the choice to simply increasecore height in Th–U EEC and Th–Pu BOL minimizes leakages, whichlimits the actinide inventory necessary to attain iso-breeding butalso impact negatively some safety parameters. To assess the im-pact of core geometry on the calculated parameters, a differentTh–U iso-breeder core has been set-up by increasing the radial-core dimension, while keeping the same lead velocity through anincreased fuel-to-lead ratio. As a result, the actinide inventory toobtain iso-breeding increased from 72 t for the iso-breeder corewith increased height to 85 t for the larger core. On the other hand,noticeable improvements can be achieved in void reactivity(��30%) and coolant expansion coefficients (��50%) of the largercore, with a negative impact only on the fuel coefficient (��10%). Aproper core optimization is beyond the scope of the present workbut it is expected to improve the coefficients in Table 5, whileslightly increasing the core actinide inventory.

In spite of the potential advantages of the thorium option at BOLand EEC, problems may arise due to their evolution during thetransition cycles from the initial core to the equilibrium cycle. Incase of U–Pu and Th–U cycles, starting from an initial fissile ofrespectively Pu and 233U, the core is expected to show featuressimilar to the equilibrium case for the entire lifetime of the reactor.In particular, the safety coefficients, primarily determined by theprevailing fertile and fissile isotopes, will show only mild varia-tions from the start-up to the equilibrium cores. On the contrary,some potentially unsafe configurations may arise as a result ofthe transition from Th–Pu BOL to the Th–U EEC state (see Table 5).To exclude this, safety parameters have been computed through-out the transition, as shown in Fig. 8a to f. Both BOC (BeginningOf Cycle) and EOC values are shown due to the notable variationwithin a cycle, especially in the first cycles. The values obtainedfor the equilibrium U–Pu and Th–U cores are also reported, at bothEEC and BEC (Beginning of Equilibrium Cycle) states. It is interest-ing to point out that the differences between EEC and BEC, for bothU–Pu and Th–U equilibrium cores, are notable only for the Dopplercoefficient and neutron generation time. For both cases, EEC repre-sents the worst condition, which a posteriori justifies its choice forthe preceding discussion and the values reported in Table 5. Voidreactivity is not reported because it features similar behavior asthe coolant coefficient.

All the safety parameters shown in Fig. 8a to f reach equilibriumvalues within 150–200 years, similarly to the case of reactivity. Ex-cept for the generation time and Doppler constant, most of thetransition happens within the first few cycles, where the variationfrom BOC to EOC can also be notable. In addition, parameters atEOC are generally closer to Th–U equilibrium than at BOC, but thisis not the case for the Doppler coefficient and generation time. Forthe Doppler coefficient, conditions at the end of first cycle are com-parable with U–Pu EEC, whereas the neutron generation time iseven worse. Both the rapid changes in the coefficients and the ini-tial worsening of Doppler and generation time could require someflexibility in the control system. The transition toward Th–U equi-librium is smooth and monotonic, confirming that the boundingvalues occur at Th–Pu BOL and Th–U equilibrium. Two exceptionsare fuel and diagrid expansion, which temporarily exceed the Th–Ucase; however, the magnitude is �0.01 pcm, comparable to the tol-erance of the present calculations. It is thus concluded that, if Th–Pu BOL and Th–U EEC states are acceptable from the safety point of

-1

-0.95

-0.9

-0.85

-0.8

-0.75

-0.7

-0.65

-0.6

-0.55

-0.5

0 50 100 150 200 250

Dop

pler

coe

ffici

ent [

pcm

/K]

Time [EFPY]BOC EOC U-PU BEC

U-PU EEC Th-U BEC Th-U EEC

U-Pu

Th-U

(a)

-0.2

-0.19

-0.18

-0.17

-0.16

-0.15-0.14

-0.13

-0.12

-0.11

-0.1

0 50 100 150 200 250

Fuel

exp

ansi

on c

oeffi

cien

t [pc

m/K

]



U-Pu

Th-U

(b)

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

0 50 100 150 200 250

Coo

lant

exp

ansi

on c

oeffi

cien

t [p

cm/K

]



U-Pu

Th-U

-0.65

-0.6

-0.55

-0.5

-0.45

-0.4

-0.35

0 50 100 150 200 250

Dia

grid

exp

ansi

on c

oeffi

cien

t [pc

m/K

]



U-Pu

Th-U

250

260

270

280

290

300

310

320

330

340

350

0 50 100 150 200 250

Beta

effe

ctiv

e [p

cm]



U-Pu

Th-U

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

1.1

1.15

1.2

0 50 100 150 200 250

Gen

erat

ion

time

[µs]



U-Pu

Th-U

(c)

(e)

(d)

(f)

Fig. 8. Evolution of safety parameters in the transition from Th–Pu to Th–U: (a) doppler coefficient, (b) fuel expansion coefficient, (c) coolant expansion coefficient, (d) diagridexpansion coefficient, (e) beta effective, and (f) generation time.


view, the transition between them should be also acceptable, witha note of caution on the relatively quick change of the parametersin the first cycle and the initial worsening of Doppler and genera-tion time.

5.2. Radio-toxicity and decay heat

This Section compares the three cores under investigation (U–Pu EEC, Th–U EEC, Th–Pu BOL) from the viewpoint of radio-toxicityand decay heat, showing the impact of Table 4 core inventory iso-

topic compositions on these fuel cycle metrics. As already dis-cussed, U–Pu EEC and Th–U EEC are equilibrium cores. Theyrepresent an asymptotic upper limit of radio-toxicity for the ura-nium and thorium fuel cycles when Pu and 233U are employed asthe predominant fissile materials for the respective start-up cores.On the other hand, when TRUs are employed for the start-up core,their radio-toxicity is the lower asymptotic limit. In the presentanalysis, this will be compared with the radio-toxicity and decayheat of Th–Pu BOL to assess the Pu-burning capabilities of the Thcore.

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Rad

io-to

xici

ty [S

v/G

We/y

r]

Decay time [years]U-Pu (Extended-EQL3D) 0.1% lossesTh-U (Extended-EQL3D) 0.1% lossesTh-Pu (Extended-EQL3D) 0.1% lossesU-Pu (ORIGEN-S) 0.1% lossesTh-U (ORIGEN-S) 0.1% lossesTh-Pu (ORIGEN-S) 0.1% lossesReference level

Th-U EEC

Th-Pu BOL

U-Pu EEC

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Rad

io-to

xici

ty [S

v/G

We/y

r]

Decay time [years]

U-Pu (Extended-EQL3D) 1% lossesTh-U (Extended-EQL3D) 1% lossesTh-Pu (Extended-EQL3D) 1% lossesU-Pu (ORIGEN-S) 1% lossesTh-U (ORIGEN-S) 1% lossesTh-Pu (ORIGEN-S) 1% lossesReference level

Th-U EEC

Th-Pu BOL

U-Pu EEC

(a)

(b)

Fig. 9. Evolution of radio-toxicities of: (a) 0.1% and (b) 1% reprocessing losses andcomparison with ORIGEN-S results and with the RL.


Radio-toxicities and decay heat as obtained with the newlyimplemented EQL3D module discussed in Section 2.2) are assessedagainst ORIGEN-S. In particular, ORIGEN-S is employed to calculatethe radio-toxicity and decay heat starting from the masses calcu-lated through the standard EQL3D procedure. A comparison be-tween ERANOS and ORIGEN-S in terms of fuel depletioncalculations is instead beyond the scope of the work.

5.2.1. Radio-toxicityIn case of FRs with full recycle of actinides, long-lived wastes

originate mainly from reprocessing and fuel manufacturing losses.Such losses will likely differ for the various actinides, fuel typesand processes. Typical values considered are below 1%, often onthe order of 0.1% (Artioli et al., 2010; NEA, 2006; Salvatores et al.,2009). In the present study, both 0.1% and 1% losses are considered.It is also common practice to adopt a Reference radio-toxicity Level(RL), often determined as the radio-toxicity of the natural uraniumrequired to fuel a typical LWR once-through core of same electricalenergy output. The natural uranium is considered to be in secularequilibrium with the progeny. The resulting RL has been calculatedas 5.9 � 106 Sv/GWe-yr, using data reported in (Rose et al., 2011)for the French EPR core, and radio-toxicity coefficients for ingestionfrom ICRP72 (1996). The RL value obtained is consistent with theresults published in NEA (2002).

5.2.1.1. U–Pu EEC, Th–U EEC and Th–Pu BOL. Fig. 9 shows the evolu-tion over time of the radio-toxicity from reprocessing losses of U–Pu EEC, Th–U EEC and Th–Pu BOL from the new EQL3D module andORIGEN-S. As mentioned, both 0.1% (Fig. 9a) and 1% (Fig. 9b) lossesare considered. Reprocessing losses and RL radio-toxicities are nor-malized per GWe-yr of energy production. Fig. 9a and b showexcellent agreement between ORIGEN-S and the extended-EQL3D.This indicates that the assumption of secular equilibrium adoptedin the development of the EQL3D radio-toxicity module leads tonegligible errors. Some problems could have been expected inthe long term radio-toxicity of Th–U EEC, which was found to bedominated by 229Th and 226Ra, two of the parent isotopes whichwere considered to be in secular equilibrium with their progeny(see Section 2.2). Nevertheless, discrepancies between ERANOSand ORIGEN-S in this period are on the order of few per cents(e.g., 1.2% at 105 years). This is explained by: (1) the time elapsedfrom the beginning of the decay, which is widely sufficient forthe onset of secular equilibrium for 229Th, 226Ra and their proge-nies; (2) the slow variation of the quantity of such isotopes, whichdoes not strongly affect the equilibrium. Besides, in the first fewyears, where secular equilibrium is not established, the contribu-tion of 226Ra, 227Ac, 228Th and 229Th to the total radio-toxicity isgenerally small, thus making the error introduced by assumingsecular equilibrium negligible. Among these four isotopes, the onlyappreciable contributor is 228Th in the Th–U case. It is a fortunatecoincidence that the longest-lived daughter of this isotope has ahalf-life of only 3.7 days (Table 1), thus allowing a rapid establish-ment of the equilibrium.

Regarding the radio-toxicity comparison of the various options,Fig. 9a and b show that Th–U EEC fosters a decreasing radio-toxic-ity during the first 1000 years, one to two orders of magnitude be-low that of U–Pu EEC. Subsequently, the radio-toxicity of Th–U EECstarts increasing. The advantages over U–Pu EEC vanish at around35,000 years. The radio-toxicity of Th–U EEC continues to increaseuntil �60,000 years, reaching a value close to that obtained after250 years, which, for the assumed range of reprocessing losses, islower than the RL. The radio-toxicity from 0.1% actinides lossesin Th–U EEC is below the RL already at the beginning of the simu-lation. In the more conservative scenario of 1% reprocessing losses,Th–U EEC radio-toxicity becomes lower than the RL line after only200 years. On the other hand, in case of U–Pu EEC, approximately

400 and 25,000 years are required for the radio-toxicity to becomelower than the RL at respectively 0.1% and 1% reprocessing losses.

These results confirm that a closed thorium cycle may representa factual option for reducing the radiotoxic burden to the finalwaste repository, with the caveat that the radio-toxicity increasesafter 1000 years and will potentially conduce to a long-term peakhigher than in the U closed cycle (but lower than the RL radio-tox-icity). There are intrinsic limitations in correlating radio-toxicitywith actual risk of exposure that should not be forgotten. In addi-tion, the RL itself, while it may represent a convenient term of com-parison, is in essence an arbitrary number and doubts may becasted on its meaningfulness. The assumed RL, �6 � 106 Sv forGWe-yr of energy produced, is arguably a reassuring level whencompared to a lethal dose of 10 Sv. Yet, it is a fact that a thoriumcycle guarantees a drastically lower radio-toxicity than the ura-nium cycle for thousands of years.

A better insight into the behaviors observed in Fig. 9a and b canbe achieved by separating the effects of the different isotopes.Fig. 10 illustrates the main contributors to the 0.1% actinide wastes

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Rad

io-to

xici

ty [S

v/G

w /

yr]

Decay time [years]

Ra226Th228Th229U232U233Pu238Total

(a)

(b)

e

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Rad

io-t

oxic

ity [S

v/G

we/y

r]

Decay time [years]

Pu238

Pu239

Pu240

Pu241

Am241

Am242m

Cm242

Cm244

Total

Fig. 10. Evolution of radio-toxicity of each isotope: (a) U–Pu EEC and (b) Th–U EEC.


radio-toxicity of U–Pu EEC (Fig. 10a) and Th–U EEC (Fig. 10b). Themajor contribution for the first few decades comes from 238Pu forboth cores. 238Pu is in fact the only TRU which is present in theTh–U EEC in a quantity roughly comparable to the U–Pu EEC (Ta-ble 4). 232U and 228Th represent the other important contributorsto radio-toxicity in the first few centuries for Th–U EEC. 228Th isthe daughter of 232U and after discharge they both have nearlyidentical activity, showing that they are in secular equilibrium.241Pu, 241Am, 242Am-m, 242Cm and 244Cm, which also contributeto radio-toxicity in the first few decades or centuries, play a roleonly in U–Pu EEC. This confirms that the very limited build-up ofTRUs in the Th–U equilibrium cycle (see Table 4) is important forreducing fuel radio-toxicity. After few centuries, radio-toxicity isdominated by 239Pu and 240Pu, in the U–Pu case, and by 229Thand 226Ra (and their progenies) in Th–U case. 229Th and 226Ra aredaughters of 233U and 234U, respectively, and the latters can actu-ally be claimed as the isotopes responsible for the radio-toxicitygrowth in Th–U EEC. The notable role of Pu isotopes explains theclose value obtained for U–Pu EEC and Th–Pu BOL. Actually, Th–Pu BOL is free from minor actinides, but features a notably higherPu inventory than U–Pu EEC (Table 4).

In Section 2, it was mentioned that equilibrium cycle radio-tox-icity can be considered as a limiting upper boundary calculationwhen the reactor operates in a fully closed fuel cycle starting froman initial core of U/Pu and Th/U main fertile/fissile materials.Fig. 10a and b confirm this assumption. In both U–Pu EEC andTh–U EEC, a major role is played by actinides which are virtuallyabsent in a fresh core, and which require many cycles for their

complete build-up. In particular, 244Cm and 238Pu notably contrib-ute to radio-toxicities of U–Pu EEC and Th–U EEC, respectively. Incase of pure plutonium and thorium cycles, i.e., without any initialfissile other than plutonium and 233U, both these isotopes are ab-sent at start-up, and require a long time for complete build-up(see Figs. 2 and 5). If a core is started with only thorium and233U, as in the case presented in Section 4, after 60 years of contin-uous operations only 16.5 kg of 238Pu are present in the core, com-pared to the 85 kg at equilibrium. Over time scales of interest forengineering applications, the actual radio-toxicity of the (pure)thorium cycle will then be considerably lower than the equilibriumone.

5.2.1.2. Transition from Th–Pu BOL to Th–U EEC. Even though Th–PuBOL and Th–U EEC have comparable or lower radio-toxicities (forthe first several thousand years) than the U–Pu case, the transitionbetween the two cores may represent an issue. In particular, theinitial Th–Pu BOL core is free of minor actinides and an initialbuild-up of radio-toxicity is expected. In addition, transition ofradio-toxicity from Th–Pu core to Th–U equilibrium is expectedto be slower than for the reactivity and safety parameters, becauseradio-toxicity may be dependent on the mass of slowly-varyingactinides like Cm and 238Pu. Fig. 11 shows the cycle-by-cycleradio-toxicity generation. The first 25 cycles are representedexplicitly, while the additional step shows the values after 100 cy-cles, which corresponds to the Th–U EEC case. The reported radio-toxicities are computed using the actinide content at BOC, whichalso represent EOC masses for the preceding cycle, after 8.5 yearsof cooling. For simplicity, only the case of 0.1% losses is considered.Fig. 11 represents an estimate of the burning capability of the Th–Pu core. In fact, a correct evaluation would first require the individ-uation of a time scale of interest. The value after this period, or theintegral over this period, would represent the correct number to beconsidered for judging the burning capabilities of a reactor (vanRooijen, 2009). Evaluations of this kind are out of the scope ofthe present work, but some conclusions can be drawn anyway withthe support of Fig. 10, which helps understanding the middle andlong term impact of each isotope. Besides, the 8.5 years coolingconsidered for Fig. 11 can be considered a time scale of interestfor reprocessing and fabrication. It can be observed that radio-tox-icity is featured by a notable increase in the first cycles, as ex-pected. For radio-toxicity to return to the original BOL values,approximately 10 cycles are required. This means that approxi-mately 80 years of continuous operation are necessary before useof thorium actually leads to a radio-toxicity reduction. Build-upof 244Cm is mainly responsible for the peak observed in Fig. 11. Thisisotope is relatively short lived (half-life equal to approximately18 years), but its daughter, 240Pu, is one of the main responsiblefor the long term radio-toxicity of the fuel. 244Cm represents alsothe main contributor to neutron emission in the first years of cool-ing (Salvatores et al., 2009). In addition, 232U (and consequently228Th) shows a rapid build-up, reaching the asymptotic value(approximately 1500 ppm of the total uranium content) in few cy-cles. 232U is particularly of concern for fuel fabrication due to thehighly penetrating gammas emitted by its progeny. After the peak,the reactor actually starts behaving as a burner, but the process isvery slow and 25 cycles are still not sufficient to reach the Th–Uequilibrium state. Transition time for the equilibrium is roughlyproportional to the inverse of the specific power (thermal powerper unit mass of the fuel) and the long transition time for the ESLYis a direct consequence of its low value for this parameter (21MWth/tonHN) compared e.g. to typical sodium fast reactors. Transi-tion can be accelerated through Pu removal by reprocessing, whatis enabled by the reactivity excess (Fig. 6), but the extracted Pushould be disposed of. The persistence in the fuel of plutonium,242Cm and 244Cm leads also to a non-negligible build-up of higher

0.0E+00

2.0E+06

4.0E+06

6.0E+06

8.0E+06

1.0E+07

1.2E+07

1.4E+07

1.6E+07

1.8E+07

Rad

io-to

xici

ty [S

v/G

W /y

r]

Cycle number [-]

Th228U232Pu238Pu239Pu240Pu241Am241Am242mCm244Others

e

Fig. 11. In-core evolution of BOC radio-toxicity in the transition from Th–Pu to Th–U equilibrium.

(a)


actinides. In particular, 246Cm, that is the main responsible for neu-tron emission in the middle term (Salvatores et al., 2009), reaches amass equal to �8 kg in the core after approximately 10 cycles. Thisvalue is very close to the 11 kg of U–Pu EEC. It may be concludedthat use of thorium with plutonium as initial fissile material isnot a promising option for radio-toxicity reduction using the ELSY,if reasonable burning time are assumed. In fact, for tens of years,short term radio-toxicity would be increased by 232U and TRUbuild-up, while the long term impact would be worsened by con-version of thorium in 233U and 234U.

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Dec

ay h

eat [

W]

Decay time [years]

U-Pu (Extended-EQL3D)Th-U (Extended-EQL3D)Th-Pu (Extended-EQL3D)U-Pu (ORIGEN-S)Th-U (ORIGEN-S)Th-Pu (ORIGEN-S)

0

1

2

3

4

5

6

7

8

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Dis

crep

ancy

[%]

Decay time [years]

U-Pu (Extended EQL3D-ORIGEN-S)Th-U (Extended EQL3D-ORIGEN-S)Th-Pu (Extended EQL3D-ORIGEN-S)

(a)

(b)

Fig. 12. Decay heat of the actinide inventory: (a) predicted evolution with time and(b) difference between extended-EQL3D and ORIGEN-S results.

5.2.2. Decay heatDecay heat produced by actinides represents a concern in a nu-

clear fuel cycle because it increases the costs of reprocessing, man-ufacturing, and, particularly, of final repository. Reprocessing andmanufacturing are performed after few years or few tens of years.Repository must be designed on the basis of the maximum heatload. If constant rate of waste loading is assumed, the maximumheat load will be reached at the final closure of the repository, withan ensuing temperature peak within at most few thousands ofyears. These considerations allow to focus the attention on a timescale from few years to few thousands of years.

Fig. 12a and b show the evolution over the time of the decayheat of U–Pu EEC, Th–U EEC and Th–Pu BOL cores, and compareit to the results computed by means of ORIGEN-S. Time stepsadopted for EQL3D and ORIGEN-S calculations are different andan interpolation of the EQL3D results has been necessary to calcu-late discrepancies in Fig. 12b, which explains the scattering in thereported data. A period up to 1 million years is still considered inorder to better assess the capabilities of the extended-EQL3D pro-cedure. In this case, set-up of a reference level is not of interest. Asa consequence, a quantification of the reprocessing losses is not re-quired and the decay heat related to the full actinide inventory hasbeen reported. Similarly to the case of radio-toxicity, decay heat inTh–U EEC is one order of magnitude lower in the first few tens ofthousands years compared to U–Pu EEC. The increase above the va-lue of U–Pu EEC and Th–Pu BOL after this period is not of great con-cern in the case of decay heat, because the repository should bedesigned for initial decay heat level. From an isotopic point of view,

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Dec

ay h

eat [

W]

Decay time [years]

Pu238

Pu239

Pu240

Pu241

Am241

Cm242

Cm244

Total

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06

Dec

ay h

eat [

W]

Decay time [years]

Ra226Th228Th229U232U233Pu238Total

(b)

Fig. 13. Evolution of decay heat of each isotope after discharge: (a) U–Pu EEC and(b) Th–U EEC.

0.E+00

1.E+04

2.E+04

3.E+04

4.E+04

5.E+04

6.E+04

7.E+04

8.E+04

9.E+04

1.E+05

Dec

ay h

eat [

W]

Cycle number [-]

Th228U232Pu238Pu239Pu240Pu241Am241Cm242Cm244Others

Fig. 14. In-core evolution of BOC decay heat in the transition from Th–Pu to Th–Uequilibrium.


a dominant role is played by the same isotopes as in the case ofradio-toxicity, but with different proportions, as can be observedin Fig. 13a and b. Like the case of radio-toxicity, 238Pu strongly con-tributes for the first few centuries for both U–Pu EEC and Th–UEEC. Main differences comes from contributions of 242Cm and232U (including 228Th), which are in this case increased. 242Cm isimportant in the first few years in U–Pu EEC, suggesting a coolingtime of at least 5 years before reprocessing and fabrication. On theother hand, 232U and its progeny are responsible for 2/3 of the totaldecay heat from actinides in Th–U EEC. As regards the transitionfrom Th–Pu BOL to equilibrium, cycle-by-cycle decay heat at BOCis reported in Fig. 14. Its value is nearly doubled in 3–4 cyclesand 15 cycles are necessary to go back to the initial value. Build-up of 244Cm is the main responsible for such behavior, but, in thiscase, also the rapid growth of 232U is particularly important.

Agreement between ORIGEN-S and extended-EQL3D is excel-lent. A systematic overestimation can be observed in Fig. 14b forthe results provided by extended-EQL3D, as a consequence of both:(1) secular equilibrium hypothesis; and (2) calculation of the heatreleased by beta decay by means of the reaction Q-value. The dis-crepancy between the two codes remains on the order of some per-cents (typically 3–4%) starting from 1 year, confirming that betadecay actually plays a minor role in decay heat of actinides (Salvat-ores, 2002). Use of Q-value for beta decay becomes instead unac-ceptable in calculating decay heat immediately after reactorshut-down. For U–Pu EEC and Th–U EEC, this approximation leadsto overestimations equal to 54% and 104% respectively. In particu-lar, immediately after shut-down decay heat from actinides isdominated by 239Np for U–Pu EEC and by 233Th and 233Pa for Th–U EEC, which are all beta emitters. 239Np accounts for 70% of thedecay heat from actinides for U–Pu EEC, while 233Th and 233Pa ac-counts for 67% and 31% of it in the Th–U EEC. Problems were notencountered for Th–Pu BOL, which is free from beta emitters atthe beginning of the decay.

6. Conclusions

In the present paper, a comparison between uranium and tho-rium fuel cycles has been carried out with reference to the iso-breeder ELSY core. Calculations have been performed employingthe ERANOS-based EQL3D procedure, which was extended for per-forming radio-toxicity and decay heat calculations. Such extensionis based on the direct simulation of a limited number of actinides(41) and by assuming secular equilibrium for the others. The re-sults with the extended-EQL3D procedure have been benchmarkedagainst ORIGEN-S results, showing an excellent agreement and

confirming the adequacy of the approximations proposed. Themain problem relates to tentatively using reaction Q-values tocompute decay heat. This overestimates heat release by beta decay,which brings to unacceptable errors immediately after fueldischarge.

The newly extended EQL3D procedure was employed to charac-terize and compare fully closed uranium and thorium cycles in theELSY core. As a first result, it was shown that a redesign of the ELSYcore with an increase of �50% in the actinide content is necessaryto make it an iso-breeder using thorium feed. Adopting the nomi-nal and modified ELSY cores, equilibrium cycle configurations havebeen achieved for both uranium and thorium feeds, starting frompure U–Pu and Th–U configurations. For completeness, also thetransition from a Th–Pu core to the Th–U equilibrium has beenanalyzed, since the thorium implementation would likely requirethe use of plutonium as initial fissile material.

As concerns the equilibrium cores, comparison between tho-rium and uranium showed clear advantages for thorium in termsof safety parameters. The diagrid expansion is the only coefficientwhich experiences a worsening compared to the uranium case.Preliminary analyses have shown that further improvements couldbe achieved through a proper core optimization. In terms of iso-tope evolution, a very limited (and extremely slow) build-up ofTRUs was observed in the thorium cycle. As a consequence,radio-toxicity and decay heat were found to be noticeably reducedcompared to the U–Pu equilibrium core in the first few thousandsof years. Assuming reprocessing losses equal to 0.1% for all theactinides, the thorium fuel cycle actinide losses would have aradio-toxicity below that of the amount of natural uranium neededto feed a typical once-through LWR. Approximately 400 years ofpost-irradiation cooling are instead necessary for the uranium cy-cle to reach the same radio-toxicity level. In case of 1% reprocessinglosses, the difference becomes even more notable and the thoriumcycle allows to reach the reference level of radio-toxicity in200 years instead of the 25,000 years of the uranium cycle. 238Puhas been found to be a main responsible for the short-termradio-toxicity and decay heat generation of the thorium option,but its build-up has been found to be particularly slow in fast-spec-trum reactors. In time scales of interest for engineering applica-tions, the short term radio-toxicity generation of the thoriumoptions is then expected to show further advantages comparedto the uranium counterpart. Advantages in the short-middle termradio-toxicity and decay heat are considered, in the authors’ view,sufficient to offset the very long-term increase of radio-toxicity ob-served in the thorium cycle.

As far as the transition from Th–Pu BOL to equilibrium is con-cerned, the obtained results indicate overall viability and sufficientreactivity throughout the transition, but also some notable limita-tions, especially from a radio-toxicity and decay heat point of view.In fact, due to the higher dimensions, the Th–Pu core is character-ized by radio-toxicity and decay heat comparable with the U–Puequilibrium already at BOL. As soon as the reactor operation isstarted, the presence of plutonium leads to a rapid build-up of244Cm, while the presence of thorium gives rise to the build-upof 232U. Both these isotopes play a major role in short-middle termradio-toxicity and decay heat, and determine their increase for asmany as 5 cycles. Due to the low specific power of ESLY, up to10–15 cycles (corresponding to 85–127.5 years) are then neededfor the radio-toxicity and decay heat to return to the original BOLvalue. Persistence of plutonium and 244Cm also fosters a notablegrowth of higher actinides, like 246Cm, which reaches a value veryclose to the one in U–Pu equilibrium. Radio-toxicity and decay heatcalculation may then suggest to avoid the Th–Pu configuration, butto start a thorium cycle in the ESLY directly with 233U, generating itby means of dedicated thorium blankets in U–Pu reactors. As con-cerns the safety parameters, transition from Th–Pu to pure Th–U


cycle resulted to be smooth and without oscillations, but the initialTh–Pu core shows a deterioration of some features with respect tothe Th–U equilibrium core. Transient analysis would be requiredfor a proper assessment, especially in view of the particularlylow value of beta effective observed for the Th–Pu at BOL. In addi-tion, the transition appeared to be quick in the first cycle, whichcould require a notable degree of flexibility in the control system.

7. Future work to be considered

The present paper has attempted a systematic comparison be-tween thorium and uranium use in the ELSY. The main objectivehas been to offer an overview of the pros and cons of the two op-tions, including in the analysis aspects related to safety, radio-tox-icity and decay heat. In view of the scoping nature of the presentedcalculations, a number of assumptions have been taken. In partic-ular, a single-batch reprocessing scheme has been assumed whileup to five-batch fuel reprocessing schemes are generally consid-ered for the ELSY. Possible future work should simulate this repro-cessing scheme, which would refine some of the results presentedin this paper. It would also allow a better understanding of issueslike initial fissile loading, reactivity swing in a cycle, and powerpeak optimization. Similarly, use of explicit FP in place of the pseu-do FP should be considered in future analysis for increased accu-racy, as well as to investigate the effect of FP on radio-toxicityand decay heat.

In case a design optimization for the Th-based core was of inter-est, these should include investigation about heterogeneous seed-and-blanket core configurations, adoption of mixed-spectrumcores, use of different fuel types (e.g., metal alloys) in the wholecore or in the blankets, and selection of an optimal specific power(accounting for effects related to 233Pa build-up) and coregeometry.

Finally, the performed analyses have singled out a particularlyslow convergence toward the equilibrium, especially for the Th cy-cle. This may be an advantage when starting the reactor directlywith 233U, since radio-toxicity generation would be lower. On theother hand, it has been shown that this greatly limits the reactorPu burning capabilities. The problem may clearly be overcomeusing e.g. sodium fast reactors or fast-spectrum molten salt reac-tors that generally feature much higher specific power, which isroughly proportional to the convergence rate to the equilibrium.Thorium use in such reactor concepts will be investigated by theauthors in future analyses.

Acknowledgments

The work presented in this paper has been performed in theframework of a collaboration between the Politecnico di Milano(Italy), the Paul Scherrer Institut (Switzerland), and WestinghouseElectric Company (USA). The authors are thankful to the Politecnic-o Di Milano and the Paul Scherrer Institut, who provided substan-tial financial support and made fully available their facilities andexpertise. The authors also wish to acknowledge the precious con-tribution of Westinghouse Electric Company that made possiblethe set-up of this collaboration framework, and steadily supportedthe development of the work. In particular, the authors wish tothank Dr. Sandro Pelloni (Paul Scherrer Institut) for his usefuland valuable suggestions throughout the development of the work.

References

Alemberti, A., Carlsson, J., Malambu, E., Orden, A., Struwe, D., Agostini, P., Monti, S.,2011. European lead fast reactor – ELSY. Nucl. Eng. Des. 241 (9), 3470–3480.

Aronson, S., Cisney, E., Gingerich, K.A., 2006. Thermal expansion of some cubicrefractory compounds of thorium. J. Am. Ceram. Soc. 50 (5), 248–252.

Artioli, C., Sarotto, M., Malambu, E., Sobolev, V., Massara, S., Ricotti, M.E., 2007. Elsy:neutronic design approach. In: Proc. Int. Conf. ENC 2007, September 16–20,Brussels, Belgium.

Artioli, C., Grasso, G., Sarotto, M., Krepel, J., 2009. ELSY neutronic analysis bydeterministic and Monte Carlo methods: an innovative concept for the controlrod systems. Proc. Int. Conf. ICAPP 2009, Tokyo, Japan, May 10–14.

Artioli, C., Grasso, G., Petrovich, C., 2010. A new paradigm for core design aimed atthe sustainability of nuclear energy: the solution of the extended equilibriumstate. Ann. Nucl. Energy 37, 915–922.

Benedict, M., Thomas, P., Levi, H.W., 1981. Nuclear Chemical Engineering, seconded. McGraw Hill.

Coates, D.J., Parks, G.Y., 2010. Actinide evolution and equilibrium in fast thoriumreactors. Ann. Nucl. Energy 371, 1076–1088.

Fiorina, C., Franceschini, F., Krepel, J., Mikityuk, K., 2011. Comparative analysis ofuranium and thorium fuel cycles in a Lead-cooled Fast Reactor from theperspective of safety and waste management. In: Proc. Int. Conf. GLOBAL 2011,December 11–16, Chiba, Japan.

Gruppelaar H., Schapira, J.P., 2006. Thorium as a waste management option.Technical Report. EU 19142 EN.

International Atomic Energy Agency, 1999. The Physical Protection of NuclearMaterial and Nuclear Facilities. Technical Report. INFCIRC/225/Rev.4.

International Atomic Energy Agency, 2002. Thorium fuel utilization: options andtrends. Technical Report. IAEA-TECDOC-1319.

International Atomic Energy Agency, 2005. Thorium fuel cycle – potential benefitsand challenges. Technical Report. IAEA-TECDOC-1450.

International Commission on Radiological Protection, 1996. ICRP Publication 72:Age-dependent Doses to the Members of the Public from Intake ofRadionuclides Part 5, Compilation of Ingestion and Inhalation Coefficients.Annals of the ICRP 26/1.

Kang, J., von Hippel, F.N., 2001. U-232 and the proliferation resistance of U-233 inspent fuel. Sci. Global Secur. 9, 1–32.

Kazimi, M.S., Czerwinski, M.J., Driscoll, M.J., Hejzlar, P., Meyer, J.E., 1999. On the useof thorium in Light Water Reactors. Tecnhical Report. MIT-NFC-TR-016.

Krane, K.S., 1988. Introductory Nuclear Physics. John Wiley & Sons, New York.Krepel, J., Pelloni, S., Mikityuk, K., Coddington, P., 2009. EQL3D: ERANOS based

equilibrium fuel cycle procedure for fast reactors. Ann. Nucl. Energy 36, 550–561.

Krepel, J., Pelloni, S., Mikityuk, K., Coddington, P., 2010. GFR equilibrium cycleanalysis with the EQL3D procedure. Nucl. Eng. Des Design 240, 905–917.

Kuegler, K., Pöppe, N., Jühe, S., Schitthelm, O., 2007. Use of thorium in the nucleartechnology – experiences in Germany. Technical Report. RWTH AachenUniversity.

MacDonald, P., Lee, C.B., 2004. Use of thoria-urania fuels in PWRs: a general reviewof a NERI project to assess feasible core designs, economics, fabricationmethods, in-pile thermal-mechanical behaviour, and waste formcharacteristics. Nucl. Technol. 147, 1–7.

Nuclear Energy Agency, 2002. Accelerator-Driven Systems (ADS) and Fast Reactors(FR) in advanced nuclear fuel cycles. Technical Report. NEA-3109.

Nuclear Energy Agency, 2006. Physics and safety of transmutation systems: a statusreport. Technical Report. NEA-6090.

Orlov, V.V., Filin, A.I., Smirnov, V.S., Sila-Novitsky, A.G., Tsykunov, V.S., Leonov, N.,Smirnov, V.P., Lopatkin, A.V., Ganev, I. Kh., Bosin, S.N., Abramob, V.A.,Emelyantseva, I., Khacharesov., G.A., Kogut, V.A., 2001. Naturally Safe Lead-Cooled Fast Reactor for Large-Scale Nuclear Power. Final Report ISTC, Moscow.

Pilarski, S., Lecarpentier, D., 2009. Neutronic assessment of liquid-metal cooled fastreactors using thorium fuel. In: Proc. Int. Conf. GLOBAL 2009, September 6–10,Paris. France.

Rimpault, G., 1997. Physics documentation of eranos the ecco cell code. TechnicalReport. RT-SPRC-LEPh-97-001.

Rimpault, G., Plisson, D., Tommasi, J., Jacqmin, R., Rieunier, J., Verrier, D., Biron, D.,2002. The ERANOS code and data system for fast reactor neutronic analyses. In:Proc. Int. Conf. PHYSOR 2002, Seoul, Korea, October 7–10.

Robertson, R. C., 1965. MSRE design and operations report – part 1 – description ofreactor design. Technical Report. ORNL-TM-0728.

Rodriguez, P., Sundaram, C.V., 1981. Nuclear and materials aspects of the thoriumfuel cycle. J. Nucl. Mater. 100, 227–249.

Rose, S.J., Wilson, J.N., Capellan, N., David, S., Guillemin, P., Ivanov, E., Meplan, O.,Nuttin, A., Siem, S., 2011. Minimization of actinide waste by multi-recycling ofthoriated fuels in the EPR reactor. Ann. Nucl. Energy 38, 2619–2624.

Ruggeri, J.M., 1999. ERANOS: manuel des methodes: reconstruction fine d’un fluxnodal. Technical Report. NT-SPRC-LEPH-99-217.

Salvatores, M., 2002. The physics of transmutation in critical or subcritical reactors.C. R. Phys. 3, 999–1012.

Salvatores, M., Chabert, C., Fazio, C., Hill, R., Peneliau, Y., Slessarev, I., Yang, W.S.,2009. Fuel cycle analysis of TRU or MA burner fast reactors with variableconversion ratio using a new algorithm at equilibrium. Nucl. Eng. Des. 239,2160–2168.

Salvatores, M., Palmiotti, G., 2011. Radioactive waste partitioning andtransmutation within advanced fuel cycles: achievements and challenges.Prog. Part. Nucl. Phys. 66, 144–166.

Sartori, A., Franceschini, F., Wenner, M., Ferroni, P., Ricotti M. E., 2011. Core PhysicsStudies and TRU Burning Potential of a Thorium-based Fuel Fast Reactor. In:Proc. 2011 ANS Winter Meeting and Nuclear Technology Expo, Washington, DC,October 30–November 3, 2011.

SCALE: A Modular Code System for Performing Standardized Computer Analyses forLicensing Evaluations, ORNL/TM-2005/39, Version 5.1, Vols. I–III, November


2006. Available from Radiation Safety Information Computational Center at OakRidge National Laboratory as CCC-732.

Sobolev, V., Malambu, E., Aït Abderrahim, H., 2009. Design of a fuel element for alead-cooled fast reactor. J. Nucl. Mater. 385, 392–399.

Sublet., J.C., Dean., C., Plisson-Rieunier, D., 2006. ECCOLIB-JEFF-3.1 libraries.Technical Report. CEA-R-6100.

Till, C. E., Chang, Y.I., Kittel, J.H., Fauske, H.K., Lineberry, M.J., Stevenson, M.G.,Amundson, P.I., Dance, K.D., 1980. Fast breeder reactor studies. TechnicalReport. Argonne National Laboratory, ANL-80-40.

Tommasi, J., 2001. Creation de pseudo-produits de fission pour eranos(bibliotheques jeff-3.1). Technical Note. SPRC/LEPh 06-201.

van Rooijen, W.F.G., 2009. Some notes on the definition of performance parametersfor reactors and fuel cycles. Nucl. Sci. Eng. 162, 299–306.

Date post:	07-Dec-2016
Category:	Documents
Upload:	marco-enrico
View:	212 times
Download:	0 times

Analysis of thorium and uranium fuel cycles in an iso-breeder lead fast reactor using extended-EQL3D...

Documents