S

Ba

b

a

ARRA

1

ewt(ain

a

0d

Nuclear Engineering and Design 241 (2011) 4794–4802

Contents lists available at ScienceDirect

Nuclear Engineering and Design

journa l homepage: www.e lsev ier .com/ locate /nucengdes

cenarios for the transmutation of actinides in CANDU reactors�

ronwyn Hylanda,∗, Brian Gihmb,1

Atomic Energy of Canada Limited, Chalk River Laboratories, Chalk River, Ontario, Canada K0J 1J0Atomic Energy of Canada Limited, 2251 Speakman Drive, Mississauga, Ontario, Canada L5K 1B2

r t i c l e i n f o

rticle history:eceived 20 November 2010eceived in revised form 27 January 2011ccepted 25 February 2011

a b s t r a c t

With world stockpiles of used nuclear fuel increasing, the need to address the long-term utilization of thisresource is being studied. Many of the transuranic (TRU) actinides in nuclear spent fuel produce decayheat for long durations, resulting in significant nuclear waste management challenges. These actinidescan be transmuted to shorter-lived isotopes to reduce the decay heat period or consumed as fuel in aCANDU(R) reactor.

Many of the design features of the CANDU reactor make it uniquely adaptable to actinide transmutation.The small, simple fuel bundle simplifies the fabrication and handling of active fuels. Online refuellingallows precise management of core reactivity and separate insertion of the actinides and fuel bundles intothe core. The high neutron economy of the CANDU reactor results in high TRU destruction to fissile-loadingratio.

This paper provides a summary of actinide transmutation schemes that have been studied in CANDUreactors at AECL, including the works performed in the past (Boczar et al., 1996; Chan et al., 1997; Hylandand Dyck, 2007; Hyland et al., 2009). The schemes studied include homogeneous scenarios in whichactinides are uniformly distributed in all fuel bundles in the reactor, as well as heterogeneous scenariosin which dedicated channels in the reactor are loaded with actinide targets and the rest of the reactor isloaded with fuel.

The transmutation schemes that are presented reflect several different partitioning schemes. Separa-

tion of americium, often with curium, from the other actinides enables targeted destruction of americium,which is a main contributor to the decay heat 100–1000 years after discharge from the reactor. Anotherscheme is group-extracted transuranic elements, in which all of the transuranic elements, plutonium(Pu), neptunium (Np), americium (Am), and curium (Cm) are extracted together and then transmuted.This paper also addresses ways of utilizing the recycled uranium, another stream from the separation ofspent nuclear fuel, in order to drive the transmutation of other actinides.

. Introduction

Nuclear energy has been an important source of electricity gen-ration in many parts of the world in the past century and itseighting in the total energy mix in the future is expected to main-

ain at least the current level, and potentially increase dramaticallyOECD/NEA, 2008). A consequence of this trend is that the gener-tion of used nuclear fuel will maintain its current pace and couldncrease significantly in the future. The increasing stockpile of used

uclear fuel introduces challenges in nuclear waste management.

Many of the transuranic (TRU) actinides in nuclear spent fuelre long-lived isotopes that produce decay heat long after they are

� CANDU® is a registered trademark of Atomic Energy of Canada Limited (AECL).∗ Corresponding author. Tel.: +1 613 584 3311; fax: +1 613 584 8198.

E-mail addresses: hylandb@aecl.ca (B. Hyland), gihmb@aecl.ca (B. Gihm).1 Tel.: +1 905 823 9040; fax: +1 905 403 7376.

029-5493/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rioi:10.1016/j.nucengdes.2011.03.039

Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.

discharged from the reactor. The time scale involved in this processis much longer than a human lifespan, leading to significant nuclearwaste management challenges.

A CANDU reactor offers attractive solutions for effectively deal-ing with used nuclear fuel from a light water reactor (LWR) fleet(Boczar et al., 1996; Chan et al., 1997; Hyland and Dyck, 2007;Hyland et al., 2009). Many of the design features of the CANDUreactor make it uniquely adaptable to actinide transmutation aswell as utilization of LWR used fuel with minimal reprocessing.The most significant feature is the high neutron economy result-ing from the heavy water moderator, which allows a high TRUdestruction rate relative to the fissile loading because more neu-trons are available for transmutation rather than being parasiticallyabsorbed in the moderator. Another important feature of a CANDU

reactor is that the refuelling is performed on-power and separatelyfor each fuel channel. This allows actinide targets to only occupydesired locations in the reactor and the residency time of the tar-gets to be adjusted separately from regular fuel bundles. Online

ghts reserved.

B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802 4795

muta

rf

tl3aCn

l

••

•

•

tw

2

ecadsfftac

R

Fig. 1. The two main trans

efuelling also allows precise management of core reactivity, andurther increases the neutron economy relative to batch refuelling.

Lastly, the small and simple fuel bundle simplifies the fabrica-ion and handling of active fuels. CANDU fuel bundles are short inength (49.52 cm) and light in weight (∼21 kg), consisting of either7 pins or 43 pins (CANFLEX®2 fuel) that simplify the fabricationnd handling of the bundles. These characteristics also enable aANDU fuel bundle to function as a target carrier with minimal oro design change to the bundle.

The actinide transmutation methods discussed in this paper areisted below:

5–60% (volume) Am/Cm loaded in the centre pin of fuel bundle,14–25 wt% Am/Cm in an inert matrix placed in 30 periphery chan-nels while the core is fuelled with recycled uranium (RU) fromreprocessed used LWR fuel at 0.9 wt% enrichment,americium mixed with low enriched uranium (LEU) loaded in thefull core, andgroup extracted TRU transmutation.

It should be noted that all scenarios presented in this paper are athe research stage and much further development and engineeringork would be required to bring these concepts into reality.

. Actinide transmutation schemes

One of the key strategies to deal with a long-lived actinide, forxample Am-241, is to transmute it (see Fig. 1). The transmutationauses decay heat generation to occur in relatively shorter periodfter the used fuel is discharged from the reactor so that monitoringuration of the waste is comparable to human lifespan. There areeveral different ways of introducing Am/Cm into a CANDU reactoror transmutation. As shown in Fig. 1, the transmutation of Am-241ollows several pathways that affect the decay heat production of

he spent fuel, and result in the production of isotopes of curiumnd plutonium. In the first step, a neutron captures onto Am-241,reating Am-242 or Am-242m.

2 CANFLEX® is a registered trademark of AECL and the Korea Atomic Energyesearch Institute (KAERI).

tion pathways of Am-241.

Several different pathways are available after the initial neutroncapture. Am-242m has a high fission cross-section, so by this paththe Am can be transmuted by fission. In the second pathway theAm-242 beta decays into Cm-242. The Cm-242 then alpha decayswith a relatively short half-life (163 days), and some of the originalamericium will end up as Pu-238. The Am-242m can also neutroncapture to Am-243, and a second neutron capture creates Am-244or Am-244m. The Am-244 nuclides both have short half-lives andbeta decay to Cm-244. Cm-244 has a relatively short half-life, andalpha decays to Pu-240. Am-242m can also decay by electron cap-ture to Pu-242. The isotopes Cm-242, Cm-244 and Pu-238 all havean impact on the decay heat of the spent fuel.

Am-242m, Cm-245, Pu-239, and Pu-241 are the fissile isotopes.The other isotopes act as a poison, capturing neutrons and reducingthe coolant void reactivity (CVR) of the bundle (if located in thecentral element). This process breeds plutonium through the pathshown in Fig. 1.

With regard to curium production, for schemes with relativelyshort irradiation time (a few years) the curium that is created isthe low-mass, short-lived curium isotopes, Cm-242 and Cm-244.These isotopes have short half-lives on the same time scale as fis-sion products, and once produced in the used fuel the curium couldbe stored and decayed similar to fission products, rather than putinto long-term storage or further transmuted.

The destruction of transuranic elements (Pu, Np, Am, Cm) isachieved through fission, resulting in fewer minor actinides requir-ing long-term waste disposal. The calculations presented here werecarried out using WIMS-AECL (Altiparmakov, 2008) with a nucleardata library with Z up to 96.

2.1. Heterogeneous method (target pin)

As part of an effort to investigate transmutation schemes thatcorrespond to different reprocessing scenarios, an investigationwas made of scenarios in which americium and curium are notseparated out from lanthanides (elements with Z = 57–71) duringreprocessing. This would greatly reduce the complexity of the par-

titioning scheme and potentially significantly reduce the cost ofreprocessing. The transmutation scheme involved this mixture ofAm/Cm/Ln being combined with an inert matrix; zirconia (ZrO2)was used in this study. This Am/Cm/Ln in zirconia material was

4796 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802

tds

dwtfy4mtWp

5tcaaendrap

atcdfiteoamTnTsFapdtbcaA5

used in the analysis was taken from the used fuel database main-tained by the Nuclear Energy Agency and is shown in Table 4 below.The data set used is Takahama-3 47.03 GWd/MT(IHE). The used fuel

Fig. 2. Design for CANFLEX fuel bundle with centre actinide target.

hen placed in the centre pin of a CANFLEX fuel bundle and irra-iated in a CANDU 6 reactor simulation. The fuel bundle design ishown in Fig. 2.

The rest of the fuel bundle was comprised of 1.0% LEU. This bun-le design allows for the transmutation of americium and curiumhile reducing the coolant void reactivity. The isotopic composi-

ion of the Am/Cm/Ln mixture is given in Table 1. This used nuclearuel is from a light water reactor (LWR), which was cooled for 10ears before reprocessing. The LWR had an initial enrichment ofwt% U-235 and an exit burnup of 50 MWd/kg initial heavy ele-ents (IHE). Only nuclides of interest to reactor physics, that is

hose with significant neutron cross-sections, are contained in theIMS-AECL library that was used for the study; other lanthanides

resent in the spent fuel have been ignored in this work.The amount of Am/Cm/Ln in the centre pin was varied between

% and 60% by volume. To obtain greater destruction of Am/Cm,he centre pin cases were designed so that the centre pin is recy-led into a fresh bundle. Demountable bundles have been in uset the National Research Universal (NRU) research reactor, locatedt the Chalk River Laboratories, for many years. This demountablelement fuel concept is employed in this study. While this tech-ology is well proven for the research reactor application, furtherevelopment would be needed to implement this concept in powereactors. After the first irradiation, the centre pin would be removednd placed in a new bundle containing fresh LEU in the remainingins.

Each successive irradiation will have a lower amount of neutronbsorber in the centre pin; thus, the exit burnup will increase andhere will be less of a reduction of the coolant void reactivity. In thisoncept the reactor would contain bundles with the centre pin atifferent irradiations (i.e., some bundles would be undergoing therst cycle, and others the second or third irradiation cycles), suchhat this effect averages out over the whole reactor. This study hasxamined four recycles of the centre pin. The exit burnup and effectn CVR are given in Table 2. The cumulative percentage change ofmericium is given in Table 3, and shown graphically in Fig. 3. Theass change of Am, Cm, Ln, and Pu per year is given in Table 3.

his number is averaged over all four cycles. Note that a negativeumber is a reduction in mass, and a positive number is an increase.hus while there is a large reduction in the mass of Am, there is amall increase in the mass of Cm, as Am does transmute to Cm (seeig. 1). There is also a small increase in the amount of lanthanides,s would be expected. The lanthanides do not fission, but are fissionroducts. Thus it is not expected that the lanthanide mass wouldecrease due to transmutation, but that more would be created dueo fission of the actinides. There is an increase in Pu mass, as woulde expected. The increase occurs primarily in the first cycle. For theases with low initial amounts of Am/Cm/Ln (5% and 10%), there isdestruction of Pu in cycles 2–4. For the cases with 15–45% initial

m/Cm/Ln there is a destruction of Pu in cycles 3 and 4, and for the0–60% initial Am/Cm/Ln there is destruction of Pu in cycle 4.

Fig. 3. Percentage change in the amount of americium as a function of initial amountof Am/Cm/Ln, for each cycle.

The percentage of the initial americium that is transmuteddecreases as the initial amount increases. After four cycles 94%and 84% of the Am is transmuted if the centre pin has 5% and60% Am/Cm/Ln initially. However, with a higher initial amount ofAm/Cm/Ln, there is more mass in the reactor, and a higher massof Am is transmuted per year. This mass transmuted is 1.2 kg/yearand 20.6 kg/year for 5% and 60% initial loadings respectively. Thechanges in exit burnup and in CVR are relatively small for successiverecycles of the centre pin. These changes, in general, increase as theinitial Am/Cm/Ln loading increases, following the trend of increas-ing mass transmuted. The 60% loading case shows an increase inburnup of 2.4 MWd/kg (IHE) between the first to fourth cycles, andan increase in CVR of 1.8 mk from the first to fourth cycles.

2.2. Heterogeneous method (target channel)

In this scenario a CANDU 6 core is fuelled with 0.9% fissile RU andperipheral channels around the outside of the reactor (30 channelstotal) are fuelled with Am/Cm in an inert matrix (Fig. 4) (Hylandet al., 2008). The isotopic composition of Am and Cm in spent fuel

Fig. 4. Schematic diagram of the Am/Cm target model of a CANDU 6 core.

B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802 4797

Table 1Isotopic composition of the americium, curium, and lanthanides in the centre pin.

Nuclide % by weight Nuclide % by weight Nuclide % by weight

Am-241 4.7294 Nd-146 6.7930 Gd-154 0.1710Am-242m 0.0077 Nd-148 3.4660 Gd-155 0.0435Am-243 1.6710 Nd-150 3.4660 Gd-156 1.0355Cm-243 0.0039 Pm-147 0.1012 Gd-157 0.0013Cm-244 0.5109 Sm-147 1.9771 Gd-158 0.2218Cm-245 0.0065 Sm-148 1.3336 Gd-160 0.0112Cm-246 0.0065 Sm-149 0.0272 Tb-159 0.0257Cm-247 0.0001 Sm-150 3.0960 Dy-160 0.0037La-139 11.4135 Sm-151 0.1054 Dy-161 0.0035Ce-140 12.0208 Sm-152 0.8458 Dy-162 0.0027Ce-142 10.4744 Sm-154 0.3637 Dy-163 0.0020Ce-144 0.0004 Eu-151 0.0086 Dy-164 0.0005Pr-141 10.4431 Eu-152 0.0002 Ho-165 0.0009Nd-142 0.2501 Eu-153 1.2015 Er-166 0.0003Nd-143 6.8181 Eu-154 0.1176Nd-144 12.8596 Eu-155 0.0127

Table 2Exit burnup of a bundle for each cycle, and the increase in CVR for cycles 2, 3, and 4 relative to cycle 1 for each initial amount of Am/Cm/Ln.

Initial % Am/Cm/Ln (by volume) Exit burnup (MWd/kg(IHE)) by cycle Change in CVR relative to first cycle (mk)

1 2 3 4 2 3 4

5 16.1 16.6 16.6 16.7 0.2 0.3 0.410 15.1 15.6 16.1 16.1 0.4 0.6 0.715 14.7 15.1 15.6 16.1 0.5 0.8 1.020 13.7 14.7 15.1 15.6 0.6 1.0 1.225 13.2 14.2 14.7 15.1 0.7 1.2 1.530 12.8 13.7 14.2 14.7 0.7 1.3 1.735 11.8 13.2 13.7 14.2 0.7 1.3 1.740 11.3 12.8 13.2 13.7 0.7 1.3 1.845 11.3 12.3 12.8 13.7 0.8 1.4 1.9

wo

mri3mimpwtm

TCA

50 10.8 11.855 10.4 11.360 9.9 10.8

as decayed for 30 years, and then the Am and Cm were separatedut. The burnup achieved for the RU is 12.2 MWd/kg(IHE).

The peripheral channels are selected for the Am/Cm trans-utation in this analysis because of their minimal effect on the

eactor operation in a CANDU 6 reactor. The portions of Am/Cmn the inert matrix fuel (IMF) mixture are 14%, 19%, 26%, and5% by weight. Silicon carbide (SiC) has been used as the IMFaterial in this model. SiC has been used in this study as it

s easy to implement into the models. The choice of the inertatrix is not important at this stage; any material that is trans-

arent to neutrons is suitable from the standpoint of the physicsork done here. If this fuel cycle were to be developed fur-

her, a study would be done to determine the most suitable inertatrix.

able 3umulative percent of americium transmuted, and the average mass of Am, Cm, and Lm/Cm/Ln.

Initial % Am/Cm/Ln (vol%) Cumulative percentage change of the amount of

1 2 3

5 −75 −87 −9210 −71 −86 −9115 −68 −85 −9020 −64 −83 −8925 −61 −81 −8830 −58 −79 −8735 −54 −76 −8540 −50 −74 −8445 −49 −72 −8350 −46 −70 −8155 −43 −67 −7960 −40 −64 −77

12.3 13.2 0.8 1.4 1.912.3 12.8 0.7 1.3 1.911.8 12.3 0.6 1.2 1.8

Full core time-average calculations have been performed forthe Am/Cm target core using RFSP version 3.04.01 to examine themaximum channel and bundle powers. The maximum channel andbundle powers for the time average case are 6660 kW and 790 kWrespectively. An instantaneous model, which generates randomages for the fuel channels, was used to analyze the power increasethat occurs when refuelling the reactor, referred to as the refuellingripple. The maximum channel and bundle powers for the refuellingripple are 7080 kW and 845 kW respectively. These values are allwithin normal CANDU reactor operating conditions.

Four different bundle designs were modelled for the Am/Cmcarrier. The bundle designs are 21-element, 24-element, and 30-element bundles and the 43-element CANFLEX bundle. Thesebundles each have a different mass of heavy elements. There is a

n transmuted per year averaged over the four cycles, for each initial amount of

americium Change in mass averaged over 4 cycles (kg/year)

4 Am Cm Ln Pu

−94 −1.2 0.2 0.3 0.1−93 −2.7 0.5 0.6 0.4−93 −4.0 0.7 0.8 0.6−92 −5.6 1.0 1.1 1.1−91 −7.2 1.3 1.4 1.6−91 −8.8 1.6 1.6 2.2−90 −10.8 2.1 1.8 3.1−89 −12.7 2.4 2.0 4.0−88 −14.3 2.7 2.1 4.9−87 −16.4 3.1 2.3 6.1−86 −18.4 3.5 2.4 7.4−84 −20.6 3.9 2.5 8.8

4798 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802

FC

lTaeC

ti4tistfahffib

otaacrtaent

Crfcd

swttamcs

Table 4Isotopic composition of the americium and curium in the central pin.

Nuclide % by weight

Am-241 85.7040Am-242m 0.0666Am-243 12.0214Cm-242 0.0002Cm-243 0.0268Cm-244 1.7555Cm-245 0.3783

for CANDU reactors, was required to compensate for the loss ofreactivity. The enrichment of the RU was also required to be var-ied by blending with SEU, in order to maintain a constant burnup

Table 5Input isotopic composition of the americium from the used nuclear fuel.

Isotope % by weight

ig. 5. Relationship between the support ratio and the transmutation of Am for theANFLEX bundle with 19% initial concentration of Am/Cm by volume.

ower mass of heavy elements in the bundles with fewer elements.he mass of heavy elements per fuel bundle are 1.06, 1.21, 1.52,nd 1.79 kg/bundle for the 21, 24, 30-element and CANFLEX (43-lement) bundles respectively for a composition of 19 wt% Am andm.

Fig. 5 shows the relationship between the support ratio (SR) andhe destruction of americium using a 30-peripheral channel load-ng scheme. A support ratio of 4 means that americium produced byGWe of LWR reactors can be loaded into 1 GWe of CANDU 6 reac-

ors. This plot is for the CANFLEX fuel bundle but the relationships the same for all of the bundle designs. A lighter fuel bundle has ahorter residence time to achieve the same percentage of Am/Cmransmutation as that for a heavier bundle. However, with a lighteruel bundle, less mass of Am/Cm can be fuelled in the reactor. Forgiven bundle with a given initial loading of Am/Cm, to obtain aigher transmutation of Am, the bundle needs to sit in the reactor

or a longer period of time. However, if the bundle is in the reactoror a longer time, then the mass throughput of Am (kg Am loadednto the reactor per year) will be lower, thus the support ratio wille lower.

The graph in Fig. 6 shows the link between the percentagef Am that can be transmuted and the support ratio, and howhe heavier bundles require a longer residence time to achieven equivalent destruction. If a high support ratio is desired, thenlower destruction rate is obtained; conversely if a high per-

ent transmutation of Am is desired then more GWe of CANDUeactors are required. Support ratios, for once-through applica-ions in fuel cycles, serve as an indication of how much minorctinides are loaded into the reactors in the various scenarios. Anffective strategy to burn Am (and other minor actinides) wouldeed to balance throughput and the actual quantity of MA that isransmuted.

A heavier bundle allows more Am to be input in to theANDU reactor at one time, but a longer irradiation time isequired to achieve the same Am/Cm destruction. Thereforeor a chosen destruction and support ratio the residence timean be chosen by selecting a lighter or heavier fuel bundleesign.

Varying the initial amount of Am/Cm in the bundle produces theame relationship as shown in Fig. 6. The amounts of Am/Cm thatere modelled are 14%, 19%, 26%, and 35% by weight. Fig. 7 shows

he effect of residence time on the destruction of Am/Cm and onhe support ratio. The calculations varying the initial concentration

re for the CANFLEX fuel bundle design only. This provides anothereans to choose the fuel design, whereby the initial Am/Cm con-

entration of the bundle can be chosen for a particular fuel bundle,upport ratio/% destruction, and residence time. If there are factors

Cm-246 0.0465Cm-247 0.0007

that put a limit on the residence time of a fuel bundle in a reactor,then choosing a lighter bundle would allow a fuel cycle option toachieve the same support ratio and transmutation of Am as with aheavier bundle.

During the irradiation, there is a growth of plutonium initiallyfrom the alpha decay of Cm-242 to Pu-238, and from the electroncapture decay of Am-242 to Pu-242. The isotopic evolution of theAm/Cm is shown in Fig. 8. The figure shows the results for the 21-element fuel bundle with an initial composition of 19% Am/Cm (byvolume). Note that for any other case, the trends are the same, butthe scales on the axes will stretch/contract.

The method of using 30-periphery channels for Am/Cm trans-mutation could be used if minimal operational change is desired.Due to lower neutron flux in the outer channels in the CANDU6 reactor, the power contribution from these channels to thetotal power is small compared to the channels in the centre ofthe core. This 30-periphery channel method will also maximizethe utilization of neutrons that would escape from the core oth-erwise. However, the trade-off between percent transmutationand support ratio can result in low actinide destruction: lessthan 10% transmutation rate per year even at low loading using21-element bundle at 19% concentration. To achieve a highertransmutation rate of ∼50%, the support ratio reduces to 3.5 forall four bundle types, that is roughly equivalent to ∼35 kg ofactinide destruction per year. Future work will include calcula-tions of the impact of these scenarios on the decay heat. Thishas not been performed at this time due to limitations in thecodes.

2.3. Homogeneous method, full core of Am/LEU

This scenario involves a different reprocessing scheme thanthose in the above cases; in this case americium only is separatedout of the used nuclear fuel. There have been some recent advancesin Am only separation (Modolo et al., 2008, 2009).

For these simulations the isotopic composition of amercium wasthe same as in Table 1. The renormalized composition is given inTable 5. The separated americium was then modelled as mixed withrecycled uranium (RU). This option utilizes the uranium from theused fuel, as well as transmuting the americium (Del Cul et al.,2009; Ellis, 2007). Americium acts as a neutron poison so extraU-235, above that in natural uranium, which is the nominal fuel

Am-241 73.8Am-242m 0.12Am-243 26.1

B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802 4799

0

10

20

30

40

50

60

70

80

90

100

0

% o

f Am

tran

smut

ed

0 21

Am d

43

21-el CANFSR 30

support ratio

destruction

65

AmFLEX Am0-el

987Residence Time (years)

24-el SR 2SR C

1110

Am1-el

CANFLEX

1312

30-elSR 2

11514 6

Am24-el

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

6 17

GW

e LW

R :

GW

e C

AN

DU

Fig. 6. The % transmutation of americium and support ratio versus residence time for each of the four different fuel bundles.

-0.5

0.5

1.5

2.5

3.5

4.5

5.5

6.5

0

10

20

30

40

50

60

70

80

90

100

191817161514131211109876543210

Supp

ort r

atio

, GW

e LW

R: G

We

CA

ND

U

% o

f Am

Tra

nsm

uted

Residence Time (years)

14% Am 19% Am 26% Am 35% Am

esiden

obc2

dfn

1

2

3

4

and 2854 g/MTIHE) although the fraction of Am transmuted is thelowest (45% and 77%) for a given burnup. The reduction in the frac-tions of Am transmuted is less significant at the higher burnup, 77%versus 79%.

Table 6Parameters for the Am/LEU.

Case Model Input amountof U-235 (wt%)

Burnup(MWd/kg(IHE))

14% SR 19% SR

Fig. 7. The % transmutation of americium and support ratio vs. r

f either 7.5 MWd/kg(IHE) or 21 MWd/kg(IHE). A CANFLEX fuelundle design was used for this study, see Fig. 2. These latticeell calculations were performed using WIMS-AECL (Altiparmakov,008).

Four different CANDU reactor cases will be discussed here. Twoifferent burnups were studied, the nominal burnup for a CANDU-6uelled with natural uranium, 7.5 MWd/kg(IHE), and a higher bur-up, 21 MWd/kg(IHE). These cases are:

. 0.28% Am, with a burnup of 7.5 MWd/kg(IHE). The centre pin ofthe bundle was 2.4% Dy in zirconia.

. 0.28% Am, with a burnup of 7.5 MWd/kg(IHE). The centre pin ofthe bundle was 7% Am in zirconia.

. 0.28% Am, with a burnup of 21 MWd/kg(IHE). The centre pin ofthe bundle was 1.3% Dy in zirconia.

. 0.28% Am, with a burnup of 21 MWd/kg(IHE). The centre pin ofthe bundle was 3.7% Am in zirconia.

26% SR 35% SR

ce time for each of the four different initial amounts of Am/Cm.

The results for the four cases are given in Tables 6–8. These fourcases have a neutron poison in the centre pin of the fuel bundlein order to lower the CVR. Cases 2 and 4, with Am in the cen-tre pin, enable the maximum amount of Am transmutation (2067

1 0.28% Am Dy 2.4% 1.01 7.52 0.28%, 7% Am in centre 1.04 7.53 0.28% Am 1.34% Dy 1.34 20.94 0.28%, 3.7% Am in centre 1.36 21.1

4800 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802

0

10

20

30

40

50

60

70

80

90

100

1514131211109876543210

% p

er In

itial

Am

ount

of A

m/C

m

Residence Time (years)Total Am + Cm + Pu Am + Cm Total AmTotal Cm Total Pu Am241Am243 Cm242 Cm244Pu238 Pu239 Pu240Pu242

Total Am + Cm + Pu

Total Am + Cm

Total Am

Am-241 Total Pu

Total Cm Pu-238

Pu-239 Pu-240Pu-242 Am-243Cm-242 Cm-244

Fig. 8. Isotopic evolution of the significant transuranic nuclides in the Am/Cm fuel for the

Table 7Results for the Am/LEU.

Case Model Input Am(g/MTIHE)

Net Am transmutation

Change (%) Masstransmuted(g/MTIHE)

1 0.28% Am Dy 2.4% 2843 −53.3 15162 0.28%, 7% Am in centre 4522 −45.7 20673 0.28% Am 1.34% Dy 2841 −79 2235

mHi(

aTttioCTcr

TS

mixed with natural uranium to form a MOX fuel. This fuel was thenirradiated in a CANDU 6 reactor. The input isotopic composition ofthe fuel is given in Table 9.

0.6

0.8

1.0

1.2

1.4

eat r

atio

for H

WR

recy

cle

cas

e 4

to re

fere

nce

cycl

e

4 0.28%, 3.7% Am in centre 3726 −77 2854

The support ratios are given in Table 8. Up to 20.7 GWe of pri-ary LWR can be supported for 1 GWe of CANDU reactor (case 2).owever, it should be noted that the fraction of Am transmuted

n this case is lower. For the higher transmutation fraction casescases 3 and 4), up to 6.1 GWe of primary PWR can be supported.

For case 4, the decay heat is reduced 40 years after discharge,nd a decrease of more than 60% is seen after 1000 years (Fig. 9).he decay heat reduction was calculated relative to two scenarios,he first scenario is a reference case with no recycling of Am intohe CANDU reactor. The second case is the case 4 recycle scenarion which Am is separated out from the PWR used fuel, with the restf the spent fuel going to a repository. The Am is recycled into theANDU reactor, and the CANDU spent fuel goes into the repository.he Am mass balance is taken into account in this calculation. Forase 4, one GWe from CANDU reactors can support 6.4 GWe of LWR

eactors.

able 8upport ratios for the Am/LEU.

Case Support ratio GWe primary PWR: GWeburner reactor

1 13.02 20.73 4.64 6.1

21-element fuel bundle with an initial composition of 19% Am/Cm by volume.

2.4. Group-extracted TRU-MOX in CANDU

In this case, a group-extracted reprocessing scheme is assumedfor the transuranic elements. All of the transuranic elements areused: Pu, Np, Am, and Cm, and their relative amounts are kept atthe same relative amounts as in the original LWR used fuel.

The input fuel simulated in this study was used nuclear fuel froma LWR that had been cooled for 30 years and then reprocessed torecover all the TRU (Forsberg et al., 2004). The cooled TRU were then

0.0

0.2

0.4

1000001000010001001010.1

Dec

ay h

Time after discharge (years)

Fig. 9. The ratio of the decay heat for the scenario case 4 to the reference case withno Am recycle.

B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 4794–4802 4801

0

20

40

60

80

100

120

403020100

% p

er in

itial

TR

U

Burnup (MWd/kg)

Total TRU Total Pu Total Am Total Np

Total TRUTotal Pu

Pu-240

Pu-239

Total AmPu-242Pu-241

Total Np Total CmPu238

during

dgcm

(p4ppaias

Tovctemt

TI

A time-average model of the core was created to examine themaximum channel and bundle powers. In addition to this model,an instantaneous snapshot model was also examined. The instanta-neous model, which generates random ages for the fuel channels,

Table 10

Total Cm Pu-238

Fig. 10. The isotopic evolution of TRU

The fuel is 4.0% TRU by weight (heavy elements). The fuel bundleesign used was the 43-element CANFLEX design. This is the sameeometry as shown in Fig. 2, but the materials are different in thisase. The centre pin has a composition of dysprosium in a zirconiaatrix in order to reduce CVR.Lattice cell calculations were performed using WIMS-AECL

Altiparmakov, 2008) and full-core modelling used the RFSP com-uter code (Rouben, 2002). The exit burnup of the fuel was3.4 MWd/kg(IHE). This gave a total TRU transmutation of 42%. Theercentage of each isotope transmuted and the mass transmuteder year are given in Table 10. Note for this table a positive value isdestruction and a negative value is a creation. There is an increase

n curium mass, but it should be noted that most curium isotopesre short-lived and contribute to the decay heat on the same timecale as the fission products.

The evolution of the transuranic elements is shown in Fig. 10.he high neutron economy of the CANDU reactor gives high valuesf destruction for the fissile nuclides. The fissile content of the TRUector, initially 63%, drops to 30%. The support ratio for this fuelycle is 11.2. This reduction in actinides corresponds to a reduc-ion of the decay heat of the used fuel by about 40% at 1000 yearsxit compared to LWR used fuel that is not reprocessed and trans-

uted in a CANDU reactor. This could have significant impact on

he capacity of a geological repository.

able 9nput isotopic composition of the transuranic nuclides.

Isotope % by weight

Np-237 4.7Pu-238 1.3Pu-239 59.2Pu-240 20.1Pu-241 3.0Pu-242 3.8Am-241 9.9Am-243 0.76Cm-243 0.001Cm-244 0.072Cm-245 0.12Cm-246 0.001

Pu-239 Pu-240

transmutation in the CANDU reactor.

Full-core calculations of a TRU-MOX core have been completedto demonstrate the feasibility of fuelling a CANDU 6 core with thisfuel. Several changes are needed with respect to the natural ura-nium (NU) fuel cycle. The fuelling scheme for NU generally involvesreplacing eight bundles in a fuel channel in each fuelling opera-tion, referred to as an eight-bundle shift. Due to the higher initialreactivity of the MOX bundle, in order not to introduce too muchlocalized reactivity into the core, the fuelling scheme is reduced to acombination of one- and two-bundle shifts, applied to 264 and 116channels respectively. A detailed study on the impact on fuellingmachine utilization has not been completed, but the higher bur-nup and therefore longer dwelling time of the fuel compensatesfor the smaller number of bundles shifted during refuelling. Thisresults in a refuelling rate of about 3 bundles per day, compared to15 bundles per day in the natural uranium CANDU 6 reactor.

The amounts of the transuranic nuclides that are transmuted in the TRU-MOXscenario.

Nuclide % Transmuted Mass transmuted (kg/year)

Np-237 51.0 18.9Total Np 48.1 17.9Pu-238 −206.6 −21.2Pu-239 78.6 348.4Pu-240 2.4 3.8Pu-241 −68.8 −16.5Pu-242 −128.2 −38.4Total Pu 41.1 276.8Am-241 84.2 65.8Am-243 −221.1 −13.3Total Am 62.2 52.4Cm-242 −6.1Cm-243 −2774.4 −0.22Cm-244 −1676.5 −9.5Cm-245 −158.6 −0.15Cm-246 −1706.0 −0.13Total Cm −2374.4 −16.1

Total TRU 41.9 330.2

4802 B. Hyland, B. Gihm / Nuclear Engineering

Table 11The maximum channel and bundle powers for the time-average and instantaneousmodels for the MOX fuel cycle.

Time-averagemodel

Instantaneousmodel

wemetav

3

iulcrat

mttCsgieubtflddusc

psmca

Maximum channel power (kW) 6300 7200Maximum bundle power (kW) 820 910

as used to analyze the power increase that occurs when refu-lling the reactor, referred to as the refuelling ripple. Values for theaximum channel and bundle powers are given in Table 11. As is

xpected, the maximum channel and bundle powers are higher forhe instantaneous model, which accounts for an increase in powers fresh fuel bundles are added to the core upon refuelling. Thesealues are all within normal CANDU reactor operating conditions.

. Conclusion

In this paper, several different actinide transmutation schemesn a CANDU reactor were presented. These methods are classifiednder two categories: heterogeneous and homogeneous actinide

oading in CANDU reactors. In the former method, actinides areonfined to target pins or bundles and loaded in the reactor sepa-ately from the rest of fuel. In the homogenous methods, actinidesre homogeneously mixed with fuel and placed in all channels inhe reactor.

The advantages of these methods are that they require mini-al operational and fuel changes in existing CANDU reactors. In

he case of using the centre pin of the bundle to carry the actinides,he concept is very similar to the “low void reactivity fuel” (LVRF)ANDU fuel design that contains a centre neutron absorber andlightly enriched uranium in the remaining fuel pins. In the hetero-eneous channel method where actinide target bundles are placedn dedicated transmutation channels of the reactor, the periph-ry channels are selected to minimize the operational effect whiletilizing neutrons in the reflector region. On-power refuelling capa-ility of a CANDU reactor allows the residence time of actinideargets to differ significantly from fuel bundles. With high neutronuence, very high destruction of the actinides can be achieved. Theisadvantage of this scheme is that the total amount of actinideestruction is limited since only small part of the reactor is beingsed for actinide transmutation. The heterogeneous methods areuitable when there are numbers of existing CANDU reactors thatan support a LWR fleet.

Two different homogeneous actinide fuels are presented in thisaper corresponding to different actinide partitioning schemes:

eparated Am mixed with LEU, and group-extracted actinidesixed with natural uranium. Significant actinide mass destruction

an be achieved using whole core loading of these homogeneousctinide fuels. The actinide transmutation in a single CANDU reactor

and Design 241 (2011) 4794–4802

is in the order of several kilograms per year, achieving the supportratio between 11 and 20 depending on the actinide-partitioningscheme. The limitation in the homogeneous method is that thefuel and targets are subject to the same neutron fluence, lower-ing the fraction of actinides destroyed during the irradiation in thecore. The discharge actinide fuel may require secondary reprocess-ing and transmutation to achieve high destruction comparable tothe heterogeneous methods. A higher fuel burnup will increase theactinide destruction fraction.

Several different strategies are being studied to maximize theCANDU reactor utilization for actinide destruction. The currentlyexisting CANDU reactors can process small quantities of actinideswithout significant operational changes. This paper has not exam-ined actinide destruction using thorium as a fuel matrix, ratherthan uranium. Thorium is anticipated to be a better transmutationmatrix than uranium because fewer higher mass actinides would beproduced during irradiation and valuable U-233 would be producedfor subsequent recycle.

References

Altiparmakov, D., 2008. New capabilities of the lattice code WIMS-AECL. In: PHYSOR-2008, International Conference on Reactor Physics, Nuclear Power: A SustainableResource , Interlaken, Switzerland.

Boczar, P.G., et al., 1996. Advanced CANDU systems for plutonium destruction. In:NATO Advanced Research Workshop on Advanced Nuclear Systems ConsumingExcess Plutonium , Moscow, Russia.

Chan, P.S.W., et al., 1997. CANDU—a versatile reactor for plutonium disposition oractinide burning. In: Global 97 Conference on Future Nuclear Systems , Yoko-hama, Japan.

Del Cul, G.D., et al., 2009. Analysis of the reuse of uranium recovered fromthe reprocessing of commercial LWR spent fuel. In: ORNL/TM-2007/207,ORNL/GNEP/LTR-2008-002 , January 2009.

Ellis, R.J., 2007. Prospects of using reprocessed uranium in CANDU reactors, in the USGNEP program. The American Nuclear Society and the European Nuclear Soci-ety 2007 International Conference on Making the Renaissance Real, November11–15, 2007, Washington, D.C. Trans. Am. Nucl. Soc. 97, 107–108.

Forsberg, C.W., et al., 2004. Can thermal reactor recycle eliminate the need formultiple repositories? In: 8th OECD/NEA Information Exchange Meeting onPartitioning and Transmutation , Las Vegas.

Hyland, B., Dyck, G.R., 2007. Actinide burning in CANDU reactors. In: Global 07Conference on Advanced Nuclear Fuel Cycles and Systems , Boise, Idaho.

Hyland, B., et al., 2008. Transmutation of actinides in CANDU reactors. In: OECD/NEAInformation Exchange Meeting on Partitioning and Transmutation , Japan.

Hyland, B., et al., 2009. Transmutation of americium in light and heavy water reac-tors. In: Global 2009 Conference on the Nuclear Fuel Cycle: Sustainable Optionsand Industrial Perspectives , Paris, France.

Modolo, G., et al., 2008. Development and demonstration of a new SANEX processfor actinide(III)/lanthanide(III) separation using a mixture of CyMe4BTBP andTODGA as selective extractant. In: Proceedings of the 10th Information ExchangeMeeting on Partitioning and Transmutation , Mito, Japan, October 2008.

Modolo, G., et al., 2009. Selective separation of americium(III) from curium(III),californium(III), and lanthanides(III) by the LUCA process. In: Global 2009Conference on the Nuclear Fuel Cycle: Sustainable Options and Industrial Per-

spectives , Paris, France.

OECD/NEA, 2008. Nuclear Energy Outlook 2008, Nuclear Energy Agency.Rouben, B., 2002. RFSP-IST, the industry standard tool computer program for CANDU

reactor core design and analysis. In: Proceedings of the 13th Pacific Basin NuclearConference , Shenzhen, China, 2002 October 21–25.