Nuclear Engineering and Design 241 (2011) 47944802

Contents lists available at ScienceDirect

Nuclear Engineering and Design
j our nal homepage: www. el sevi er . com/ l ocat e/ nucengdes
Scenarios for the transmutation of actinides in CANDU reactors

Bronwyn Hyland
a,
, Brian Gihm
b,1
a
Atomic Energy of Canada Limited, Chalk River Laboratories, Chalk River, Ontario, Canada K0J 1J0
b
Atomic Energy of Canada Limited, 2251 Speakman Drive, Mississauga, Ontario, Canada L5K 1B2
a r t i c l e i n f o
Article history:
Received 20 November 2010
Received in revised form 27 January 2011
Accepted 25 February 2011
a b s t r a c t
Withworld stockpiles of used nuclear fuel increasing, the need to address the long-termutilizationof this
resource is being studied. Many of the transuranic (TRU) actinides in nuclear spent fuel produce decay
heat for long durations, resulting in signicant nuclear waste management challenges. These actinides
can be transmuted to shorter-lived isotopes to reduce the decay heat period or consumed as fuel in a
CANDU(R) reactor.
Manyof thedesignfeatures of theCANDUreactor makeit uniquelyadaptabletoactinidetransmutation.
The small, simple fuel bundle simplies the fabrication and handling of active fuels. Online refuelling
allows precise management of core reactivity andseparate insertionof the actinides andfuel bundles into
thecore. Thehighneutroneconomyof theCANDUreactor results inhighTRUdestructiontossile-loading
ratio.
This paper provides a summary of actinide transmutation schemes that have been studied in CANDU
reactors at AECL, including the works performed in the past (Boczar et al., 1996; Chan et al., 1997; Hyland
and Dyck, 2007; Hyland et al., 2009). The schemes studied include homogeneous scenarios in which
actinides are uniformly distributed in all fuel bundles in the reactor, as well as heterogeneous scenarios
in which dedicated channels in the reactor are loaded with actinide targets and the rest of the reactor is
loaded with fuel.
The transmutation schemes that are presented reect several different partitioning schemes. Separa-
tionof americium, oftenwithcurium, fromthe other actinides enables targeteddestructionof americium,
which is a main contributor to the decay heat 1001000 years after discharge from the reactor. Another
scheme 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 of
spent nuclear fuel, in order to drive the transmutation of other actinides.
Crown Copyright 2011 Published by Elsevier B.V. All rights reserved.
1. Introduction
Nuclear energy has been an important source of electricity gen-
eration in many parts of the world in the past century and its
weighting in the total energy mix in the future is expected to main-
tain at least the current level, and potentially increase dramatically
(OECD/NEA, 2008). A consequence of this trend is that the gener-
ation of used nuclear fuel will maintain its current pace and could
increase signicantly in the future. The increasing stockpile of used
nuclear fuel introduces challenges in nuclear waste management.
Many of the transuranic (TRU) actinides in nuclear spent fuel
are 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: [email protected] (B. Hyland), [email protected] (B. Gihm).
1
Tel.: +1 905 823 9040; fax: +1 905 403 7376.
discharged fromthe reactor. The time scale involved in this process
is muchlonger thana humanlifespan, leading to signicant nuclear
waste management challenges.
A CANDU reactor offers attractive solutions for effectively deal-
ing with used nuclear fuel from a light water reactor (LWR) eet
(Boczar et al., 1996; Chan et al., 1997; Hyland and Dyck, 2007;
Hyland et al., 2009). Many of the design features of the CANDU
reactor make it uniquely adaptable to actinide transmutation as
well as utilization of LWR used fuel with minimal reprocessing.
The most signicant feature is the high neutron economy result-
ing from the heavy water moderator, which allows a high TRU
destruction rate relative to the ssile loading because more neu-
trons are available for transmutationrather thanbeing parasitically
absorbed in the moderator. Another important feature of a CANDU
reactor is that the refuelling is performed on-power and separately
for each fuel channel. This allows actinide targets to only occupy
desired locations in the reactor and the residency time of the tar-
gets to be adjusted separately from regular fuel bundles. Online
0029-5493/$ see front matter. Crown Copyright 2011 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.nucengdes.2011.03.039
B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802 4795
Fig. 1. The two main transmutation pathways of Am-241.
refuelling also allows precise management of core reactivity, and
further increases the neutron economy relative to batch refuelling.
Lastly, the small and simple fuel bundle simplies the fabrica-
tion and handling of active fuels. CANDU fuel bundles are short in
length (49.52cm) and light in weight (21kg), consisting of either
37 pins or 43 pins (CANFLEX
2
fuel) that simplify the fabrication
and handling of the bundles. These characteristics also enable a
CANDU fuel bundle to function as a target carrier with minimal or
no design change to the bundle.
The actinide transmutation methods discussed in this paper are
listed below:

560% (volume) Am/Cm loaded in the centre pin of fuel bundle,

1425wt%Am/Cminaninert matrixplacedin30peripherychan-
nels while the core is fuelled with recycled uranium (RU) from
reprocessed used LWR fuel at 0.9wt% enrichment,

americiummixed with lowenriched uranium(LEU) loaded in the

full core, and

group extracted TRU transmutation.

It shouldbenotedthat all scenarios presentedinthis paper areat
the research stage and much further development and engineering
work would be required to bring these concepts into reality.
2. Actinide transmutation schemes
One of the key strategies to deal with a long-lived actinide, for
example Am-241, is to transmute it (see Fig. 1). The transmutation
causes decay heat generation to occur in relatively shorter period
after the usedfuel is dischargedfromthe reactor so that monitoring
duration of the waste is comparable to human lifespan. There are
several different ways of introducing Am/Cminto a CANDUreactor
for transmutation. As shown in Fig. 1, the transmutation of Am-241
follows several pathways that affect the decay heat production of
the spent fuel, and result in the production of isotopes of curium
and plutonium. In the rst step, a neutron captures onto Am-241,
creating Am-242 or Am-242m.
2
CANFLEX

is a registered trademark of AECL and the Korea Atomic Energy

Research Institute (KAERI).
Several different pathways are available after the initial neutron
capture. Am-242m has a high ssion cross-section, so by this path
the Am can be transmuted by ssion. In the second pathway the
Am-242 beta decays into Cm-242. The Cm-242 then alpha decays
with a relatively short half-life (163 days), and some of the original
americium will end up as Pu-238. The Am-242m can also neutron
capture to Am-243, and a second neutron capture creates Am-244
or Am-244m. The Am-244 nuclides both have short half-lives and
beta decay to Cm-244. Cm-244 has a relatively short half-life, and
alpha 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 have
an impact on the decay heat of the spent fuel.
Am-242m, Cm-245, Pu-239, and Pu-241 are the ssile isotopes.
The other isotopes act as a poison, capturing neutrons and reducing
the coolant void reactivity (CVR) of the bundle (if located in the
central element). This process breeds plutonium through the path
shown in Fig. 1.
With regard to curium production, for schemes with relatively
short irradiation time (a few years) the curium that is created is
the low-mass, short-lived curium isotopes, Cm-242 and Cm-244.
These isotopes have short half-lives on the same time scale as s-
sion products, and once produced in the used fuel the curiumcould
be stored and decayed similar to ssion products, rather than put
into long-term storage or further transmuted.
The destruction of transuranic elements (Pu, Np, Am, Cm) is
achieved through ssion, resulting in fewer minor actinides requir-
ing long-termwaste disposal. The calculations presentedhere were
carried out using WIMS-AECL (Altiparmakov, 2008) with a nuclear
data library with Z up to 96.
2.1. Heterogeneous method (target pin)
As part of an effort to investigate transmutation schemes that
correspond to different reprocessing scenarios, an investigation
was made of scenarios in which americium and curium are not
separated out from lanthanides (elements with Z =5771) during
reprocessing. This would greatly reduce the complexity of the par-
titioning scheme and potentially signicantly reduce the cost of
reprocessing. The transmutation scheme involved this mixture of
Am/Cm/Ln being combined with an inert matrix; zirconia (ZrO
2
)
was used in this study. This Am/Cm/Ln in zirconia material was
4796 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802
Fig. 2. Design for CANFLEX fuel bundle with centre actinide target.
then placed in the centre pin of a CANFLEX fuel bundle and irra-
diated in a CANDU 6 reactor simulation. The fuel bundle design is
shown in Fig. 2.
The rest of the fuel bundle was comprised of 1.0% LEU. This bun-
dle design allows for the transmutation of americium and curium
while reducing the coolant void reactivity. The isotopic composi-
tion of the Am/Cm/Ln mixture is given in Table 1. This used nuclear
fuel is from a light water reactor (LWR), which was cooled for 10
years before reprocessing. The LWR had an initial enrichment of
4wt% U-235 and an exit burnup of 50MWd/kg initial heavy ele-
ments (IHE). Only nuclides of interest to reactor physics, that is
those with signicant neutron cross-sections, are contained in the
WIMS-AECL library that was used for the study; other lanthanides
present in the spent fuel have been ignored in this work.
The amount of Am/Cm/Ln in the centre pin was varied between
5% and 60% by volume. To obtain greater destruction of Am/Cm,
the centre pin cases were designed so that the centre pin is recy-
cled into a fresh bundle. Demountable bundles have been in use
at the National Research Universal (NRU) research reactor, located
at the Chalk River Laboratories, for many years. This demountable
element fuel concept is employed in this study. While this tech-
nology is well proven for the research reactor application, further
development would be needed to implement this concept inpower
reactors. After therst irradiation, thecentrepinwouldberemoved
and placed in a new bundle containing fresh LEU in the remaining
pins.
Each successive irradiation will have a lower amount of neutron
absorber in the centre pin; thus, the exit burnup will increase and
there will be less of a reduction of the coolant void reactivity. In this
concept the reactor would contain bundles with the centre pin at
different irradiations (i.e., some bundles would be undergoing the
rst cycle, and others the second or third irradiation cycles), such
that this effect averages out over the whole reactor. This study has
examined four recycles of the centre pin. The exit burnup and effect
on CVR are given in Table 2. The cumulative percentage change of
americium is given in Table 3, and shown graphically in Fig. 3. The
mass change of Am, Cm, Ln, and Pu per year is given in Table 3.
This number is averaged over all four cycles. Note that a negative
number is a reductioninmass, anda positive number is anincrease.
Thus while there is a large reduction in the mass of Am, there is a
small increase in the mass of Cm, as Amdoes transmute to Cm(see
Fig. 1). There is also a small increase in the amount of lanthanides,
as wouldbe expected. The lanthanides do not ssion, but are ssion
products. Thus it is not expected that the lanthanide mass would
decrease due to transmutation, but that more wouldbe createddue
to ssion of the actinides. There is an increase in Pu mass, as would
be expected. The increase occurs primarily in the rst cycle. For the
cases with low initial amounts of Am/Cm/Ln (5% and 10%), there is
a destruction of Pu in cycles 24. For the cases with 1545% initial
Am/Cm/Ln there is a destruction of Pu in cycles 3 and 4, and for the
5060% initial Am/Cm/Ln there is destruction of Pu in cycle 4.
Fig. 3. Percentage change inthe amount of americiumas a functionof initial amount
of Am/Cm/Ln, for each cycle.
The percentage of the initial americium that is transmuted
decreases as the initial amount increases. After four cycles 94%
and 84% of the Am is transmuted if the centre pin has 5% and
60% Am/Cm/Ln initially. However, with a higher initial amount of
Am/Cm/Ln, there is more mass in the reactor, and a higher mass
of Am is transmuted per year. This mass transmuted is 1.2kg/year
and 20.6kg/year for 5% and 60% initial loadings respectively. The
changes inexit burnupandinCVRarerelativelysmall for successive
recycles of the centre pin. These changes, in general, increase as the
initial Am/Cm/Ln loading increases, following the trend of increas-
ing mass transmuted. The 60% loading case shows an increase in
burnup of 2.4MWd/kg (IHE) between the rst to fourth cycles, and
an increase in CVR of 1.8mk from the rst to fourth cycles.
2.2. Heterogeneous method (target channel)
Inthis scenarioa CANDU6core is fuelledwith0.9%ssile RUand
peripheral channels around the outside of the reactor (30 channels
total) are fuelled with Am/Cm in an inert matrix (Fig. 4) (Hyland
et al., 2008). The isotopic composition of Am and Cm in spent fuel
used in the analysis was taken from the used fuel database main-
tainedby the Nuclear Energy Agency andis showninTable 4 below.
The data set usedis Takahama-3 47.03GWd/MT(IHE). The usedfuel
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) 47944802 4797
Table 1
Isotopic 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 0.1710
Am-242m 0.0077 Nd-148 3.4660 Gd-155 0.0435
Am-243 1.6710 Nd-150 3.4660 Gd-156 1.0355
Cm-243 0.0039 Pm-147 0.1012 Gd-157 0.0013
Cm-244 0.5109 Sm-147 1.9771 Gd-158 0.2218
Cm-245 0.0065 Sm-148 1.3336 Gd-160 0.0112
Cm-246 0.0065 Sm-149 0.0272 Tb-159 0.0257
Cm-247 0.0001 Sm-150 3.0960 Dy-160 0.0037
La-139 11.4135 Sm-151 0.1054 Dy-161 0.0035
Ce-140 12.0208 Sm-152 0.8458 Dy-162 0.0027
Ce-142 10.4744 Sm-154 0.3637 Dy-163 0.0020
Ce-144 0.0004 Eu-151 0.0086 Dy-164 0.0005
Pr-141 10.4431 Eu-152 0.0002 Ho-165 0.0009
Nd-142 0.2501 Eu-153 1.2015 Er-166 0.0003
Nd-143 6.8181 Eu-154 0.1176
Nd-144 12.8596 Eu-155 0.0127
Table 2
Exit 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 rst cycle (mk)
1 2 3 4 2 3 4
5 16.1 16.6 16.6 16.7 0.2 0.3 0.4
10 15.1 15.6 16.1 16.1 0.4 0.6 0.7
15 14.7 15.1 15.6 16.1 0.5 0.8 1.0
20 13.7 14.7 15.1 15.6 0.6 1.0 1.2
25 13.2 14.2 14.7 15.1 0.7 1.2 1.5
30 12.8 13.7 14.2 14.7 0.7 1.3 1.7
35 11.8 13.2 13.7 14.2 0.7 1.3 1.7
40 11.3 12.8 13.2 13.7 0.7 1.3 1.8
45 11.3 12.3 12.8 13.7 0.8 1.4 1.9
50 10.8 11.8 12.3 13.2 0.8 1.4 1.9
55 10.4 11.3 12.3 12.8 0.7 1.3 1.9
60 9.9 10.8 11.8 12.3 0.6 1.2 1.8
was decayed for 30 years, and then the Amand Cmwere separated
out. The burnup achieved for the RU is 12.2MWd/kg(IHE).
The peripheral channels are selected for the Am/Cm trans-
mutation in this analysis because of their minimal effect on the
reactor operation in a CANDU 6 reactor. The portions of Am/Cm
in the inert matrix fuel (IMF) mixture are 14%, 19%, 26%, and
35% by weight. Silicon carbide (SiC) has been used as the IMF
material in this model. SiC has been used in this study as it
is easy to implement into the models. The choice of the inert
matrix is not important at this stage; any material that is trans-
parent to neutrons is suitable from the standpoint of the physics
work done here. If this fuel cycle were to be developed fur-
ther, a study would be done to determine the most suitable inert
matrix.
Full core time-average calculations have been performed for
the Am/Cm target core using RFSP version 3.04.01 to examine the
maximumchannel and bundle powers. The maximumchannel and
bundle powers for the time average case are 6660kW and 790kW
respectively. An instantaneous model, which generates random
ages for the fuel channels, was used to analyze the power increase
that occurs when refuelling the reactor, referred to as the refuelling
ripple. The maximumchannel and bundle powers for the refuelling
ripple are 7080kW and 845kW respectively. These values are all
within normal CANDU reactor operating conditions.
Four different bundle designs were modelled for the Am/Cm
carrier. The bundle designs are 21-element, 24-element, and 30-
element bundles and the 43-element CANFLEX bundle. These
bundles each have a different mass of heavy elements. There is a
Table 3
Cumulative percent of americium transmuted, and the average mass of Am, Cm, and Ln transmuted per year averaged over the four cycles, for each initial amount of
Am/Cm/Ln.
Initial % Am/Cm/Ln (vol%) Cumulative percentage change of the amount of americium Change in mass averaged over 4 cycles (kg/year)
1 2 3 4 Am Cm Ln Pu
5 75 87 92 94 1.2 0.2 0.3 0.1
10 71 86 91 93 2.7 0.5 0.6 0.4
15 68 85 90 93 4.0 0.7 0.8 0.6
20 64 83 89 92 5.6 1.0 1.1 1.1
25 61 81 88 91 7.2 1.3 1.4 1.6
30 58 79 87 91 8.8 1.6 1.6 2.2
35 54 76 85 90 10.8 2.1 1.8 3.1
40 50 74 84 89 12.7 2.4 2.0 4.0
45 49 72 83 88 14.3 2.7 2.1 4.9
50 46 70 81 87 16.4 3.1 2.3 6.1
55 43 67 79 86 18.4 3.5 2.4 7.4
60 40 64 77 84 20.6 3.9 2.5 8.8
4798 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802
Fig. 5. Relationship between the support ratio and the transmutation of Amfor the
CANFLEX bundle with 19% initial concentration of Am/Cm by volume.
lower mass of heavy elements in the bundles with fewer elements.
The mass of heavy elements per fuel bundle are 1.06, 1.21, 1.52,
and 1.79kg/bundle for the 21, 24, 30-element and CANFLEX (43-
element) bundles respectively for a composition of 19wt% Amand
Cm.
Fig. 5 shows the relationship between the support ratio (SR) and
the destruction of americium using a 30-peripheral channel load-
ing scheme. Asupport ratio of 4means that americiumproducedby
4GWe of LWR reactors can be loaded into 1GWe of CANDU 6 reac-
tors. This plot is for the CANFLEX fuel bundle but the relationship
is the same for all of the bundle designs. A lighter fuel bundle has a
shorter residence time to achieve the same percentage of Am/Cm
transmutation as that for a heavier bundle. However, with a lighter
fuel bundle, less mass of Am/Cm can be fuelled in the reactor. For
a given bundle with a given initial loading of Am/Cm, to obtain a
higher transmutation of Am, the bundle needs to sit in the reactor
for a longer period of time. However, if the bundle is in the reactor
for a longer time, then the mass throughput of Am (kg Am loaded
into the reactor per year) will be lower, thus the support ratio will
be lower.
The graph in Fig. 6 shows the link between the percentage
of Am that can be transmuted and the support ratio, and how
the heavier bundles require a longer residence time to achieve
an equivalent destruction. If a high support ratio is desired, then
a lower destruction rate is obtained; conversely if a high per-
cent transmutation of Am is desired then more GWe of CANDU
reactors are required. Support ratios, for once-through applica-
tions in fuel cycles, serve as an indication of how much minor
actinides are loaded into the reactors in the various scenarios. An
effective strategy to burn Am (and other minor actinides) would
need to balance throughput and the actual quantity of MA that is
transmuted.
A heavier bundle allows more Am to be input in to the
CANDU reactor at one time, but a longer irradiation time is
required to achieve the same Am/Cm destruction. Therefore
for a chosen destruction and support ratio the residence time
can be chosen by selecting a lighter or heavier fuel bundle
design.
Varying the initial amount of Am/Cminthe bundle produces the
same relationship as shown in Fig. 6. The amounts of Am/Cm that
were modelled are 14%, 19%, 26%, and 35% by weight. Fig. 7 shows
the effect of residence time on the destruction of Am/Cm and on
the support ratio. The calculations varying the initial concentration
are for the CANFLEX fuel bundle design only. This provides another
means to choose the fuel design, whereby the initial Am/Cm con-
centration of the bundle can be chosen for a particular fuel bundle,
support ratio/% destruction, and residence time. If there are factors
Table 4
Isotopic composition of the americium and curium in the central pin.
Nuclide % by weight
Am-241 85.7040
Am-242m 0.0666
Am-243 12.0214
Cm-242 0.0002
Cm-243 0.0268
Cm-244 1.7555
Cm-245 0.3783
Cm-246 0.0465
Cm-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 to
achieve the same support ratio and transmutation of Am as with a
heavier bundle.
During the irradiation, there is a growth of plutonium initially
from the alpha decay of Cm-242 to Pu-238, and from the electron
capture decay of Am-242 to Pu-242. The isotopic evolution of the
Am/Cm is shown in Fig. 8. The gure shows the results for the 21-
element fuel bundle with an initial composition of 19% Am/Cm(by
volume). Note that for any other case, the trends are the same, but
the 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 ux in the outer channels in the CANDU
6 reactor, the power contribution from these channels to the
total power is small compared to the channels in the centre of
the core. This 30-periphery channel method will also maximize
the utilization of neutrons that would escape from the core oth-
erwise. However, the trade-off between percent transmutation
and support ratio can result in low actinide destruction: less
than 10% transmutation rate per year even at low loading using
21-element bundle at 19% concentration. To achieve a higher
transmutation rate of 50%, the support ratio reduces to 3.5 for
all four bundle types, that is roughly equivalent to 35kg of
actinide destruction per year. Future work will include calcula-
tions of the impact of these scenarios on the decay heat. This
has not been performed at this time due to limitations in the
codes.
2.3. Homogeneous method, full core of Am/LEU
This scenario involves a different reprocessing scheme than
those in the above cases; in this case americium only is separated
out of the used nuclear fuel. There have been some recent advances
in Am only separation (Modolo et al., 2008, 2009).
For these simulations the isotopic compositionof amerciumwas
the same as in Table 1. The renormalized composition is given in
Table 5. The separatedamericiumwas thenmodelledas mixedwith
recycled uranium (RU). This option utilizes the uranium from the
used fuel, as well as transmuting the americium (Del Cul et al.,
2009; Ellis, 2007). Americium acts as a neutron poison so extra
U-235, above that in natural uranium, which is the nominal fuel
for CANDU reactors, was required to compensate for the loss of
reactivity. The enrichment of the RU was also required to be var-
ied by blending with SEU, in order to maintain a constant burnup
Table 5
Input isotopic composition of the americium from the used nuclear fuel.
Isotope % by weight
Am-241 73.8
Am-242m 0.12
Am-243 26.1
B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802 4799
0
10
20
30
40
50
60
70
80
90
100
0
%

o
f

A
m

t
r
a
n
s
m
u
t
e
d
0 2 1
Am d
4 3
21-el
CANF
SR 30
support ratio
destruction
6 5
Am
FLEX Am
0-el
9 8 7
Residence Time (years)
24-el
SR 2
SR C
11 10
Am
1-el
CANFLEX
13 12
30-el
SR 2
1 15 14 6
Am
24-el
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
6 17
G
W
e

L
W
R

:

G
W
e

C
A
N
D
U
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
19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
S
u
p
p
o
r
t

r
a
t
i
o
,

G
W
e

L
W
R
:

G
W
e

C
A
N
D
U
%

o
f

A
m

T
r
a
n
s
m
u
t
e
d
Residence Time (years)
14% Am 19% Am 26% Am 35% Am
14% SR 19% SR 26% SR 35% SR
Fig. 7. The % transmutation of americium and support ratio vs. residence time for each of the four different initial amounts of Am/Cm.
of either 7.5MWd/kg(IHE) or 21MWd/kg(IHE). A CANFLEX fuel
bundle design was used for this study, see Fig. 2. These lattice
cell calculations were performed using WIMS-AECL (Altiparmakov,
2008).
Four different CANDU reactor cases will be discussed here. Two
different burnups were studied, the nominal burnupfor a CANDU-6
fuelled with natural uranium, 7.5MWd/kg(IHE), and a higher bur-
nup, 21MWd/kg(IHE). These cases are:
1. 0.28% Am, with a burnup of 7.5MWd/kg(IHE). The centre pin of
the bundle was 2.4% Dy in zirconia.
2. 0.28% Am, with a burnup of 7.5MWd/kg(IHE). The centre pin of
the bundle was 7% Am in zirconia.
3. 0.28% Am, with a burnup of 21MWd/kg(IHE). The centre pin of
the bundle was 1.3% Dy in zirconia.
4. 0.28% Am, with a burnup of 21MWd/kg(IHE). The centre pin of
the bundle was 3.7% Am in zirconia.
The results for the four cases are given in Tables 68. These four
cases have a neutron poison in the centre pin of the fuel bundle
in order to lower the CVR. Cases 2 and 4, with Am in the cen-
tre pin, enable the maximum amount of Am transmutation (2067
and 2854g/MTIHE) although the fraction of Am transmuted is the
lowest (45% and 77%) for a given burnup. The reduction in the frac-
tions of Amtransmuted is less signicant at the higher burnup, 77%
versus 79%.
Table 6
Parameters for the Am/LEU.
Case Model Input amount
of U-235 (wt%)
Burnup
(MWd/kg(IHE))
1 0.28% Am Dy 2.4% 1.01 7.5
2 0.28%, 7% Am in centre 1.04 7.5
3 0.28% Am 1.34% Dy 1.34 20.9
4 0.28%, 3.7% Am in centre 1.36 21.1
4800 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802

0
10
20
30
40
50
60
70
80
90
100
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
%

p
e
r

I
n
i
t
i
a
l

A
m
o
u
n
t

o
f

A
m
/
C
m
Residence Time (years)
Total Am + Cm + Pu Am + Cm Total Am
Total Cm Total Pu Am241
Am243 Cm242 Cm244
Pu238 Pu239 Pu240
Pu242
Total Am + Cm + Pu
Total Am + Cm
Total Am
Am-241
Total Pu
Total Cm
Pu-238
Pu-239
Pu-240
Pu-242 Am-243
Cm-242
Cm-244
Fig. 8. Isotopic evolution of the signicant transuranic nuclides in the Am/Cm fuel for the 21-element fuel bundle with an initial composition of 19% Am/Cm by volume.
Table 7
Results for the Am/LEU.
Case Model Input Am
(g/MTIHE)
Net Am transmutation
Change (%) Mass
transmuted
(g/MTIHE)
1 0.28% Am Dy 2.4% 2843 53.3 1516
2 0.28%, 7% Am in centre 4522 45.7 2067
3 0.28% Am 1.34% Dy 2841 79 2235
4 0.28%, 3.7% Am in centre 3726 77 2854
The support ratios are given in Table 8. Up to 20.7GWe of pri-
mary LWR can be supported for 1GWe of CANDU reactor (case 2).
However, it should be noted that the fraction of Am transmuted
in this case is lower. For the higher transmutation fraction cases
(cases 3 and 4), up to 6.1GWe of primary PWR can be supported.
For case 4, the decay heat is reduced 40 years after discharge,
and a decrease of more than 60% is seen after 1000 years (Fig. 9).
The decay heat reduction was calculated relative to two scenarios,
the rst scenario is a reference case with no recycling of Am into
the CANDU reactor. The second case is the case 4 recycle scenario
in which Amis separated out fromthe PWR used fuel, with the rest
of the spent fuel going to a repository. The Am is recycled into the
CANDUreactor, and the CANDUspent fuel goes into the repository.
The Am mass balance is taken into account in this calculation. For
case 4, one GWe fromCANDUreactors can support 6.4GWe of LWR
reactors.
Table 8
Support ratios for the Am/LEU.
Case Support ratio GWe primary PWR: GWe
burner reactor
1 13.0
2 20.7
3 4.6
4 6.1
2.4. Group-extracted TRU-MOX in CANDU
In this case, a group-extracted reprocessing scheme is assumed
for the transuranic elements. All of the transuranic elements are
used: Pu, Np, Am, and Cm, and their relative amounts are kept at
the same relative amounts as in the original LWR used fuel.
The input fuel simulatedinthis studywas usednuclear fuel from
a LWR that had been cooled for 30 years and then reprocessed to
recover all theTRU(Forsberget al., 2004). ThecooledTRUwerethen
mixed with natural uraniumto forma MOX fuel. This fuel was then
irradiated in a CANDU 6 reactor. The input isotopic composition of
the fuel is given in Table 9.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
100000 10000 1000 100 10 1 0.1
D
e
c
a
y

h
e
a
t

r
a
t
i
o

f
o
r

H
W
R

r
e
c
y
c
l
e


c
a
s
e

4

t
o

r
e
f
e
r
e
n
c
e

c
y
c
l
e
Time after discharge (years)
Fig. 9. The ratio of the decay heat for the scenario case 4 to the reference case with
no Am recycle.
B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802 4801
0
20
40
60
80
100
120
40 30 20 10 0
%

p
e
r

i
n
i
t
i
a
l

T
R
U
Burnup (MWd/kg)
Total TRU Total Pu Total Am Total Np
Total Cm Pu-238 Pu-239 Pu-240
Total TRU
Total Pu
Pu-240
Pu-239
Total Am
Pu-242
Pu-241
Total Np
Total Cm
Pu238
Fig. 10. The isotopic evolution of TRU during transmutation in the CANDU reactor.
The fuel is 4.0%TRUby weight (heavy elements). The fuel bundle
design used was the 43-element CANFLEX design. This is the same
geometry as shown in Fig. 2, but the materials are different in this
case. The centre pin has a composition of dysprosium in a zirconia
matrix in order to reduce CVR.
Lattice cell calculations were performed using WIMS-AECL
(Altiparmakov, 2008) and full-core modelling used the RFSP com-
puter code (Rouben, 2002). The exit burnup of the fuel was
43.4MWd/kg(IHE). This gave a total TRUtransmutation of 42%. The
percentage of each isotope transmuted and the mass transmuted
per year are given in Table 10. Note for this table a positive value is
a destruction and a negative value is a creation. There is an increase
in curium mass, but it should be noted that most curium isotopes
are short-lived and contribute to the decay heat on the same time
scale as the ssion products.
The evolution of the transuranic elements is shown in Fig. 10.
The high neutron economy of the CANDU reactor gives high values
of destruction for the ssile nuclides. The ssile content of the TRU
vector, initially 63%, drops to 30%. The support ratio for this fuel
cycle is 11.2. This reduction in actinides corresponds to a reduc-
tion of the decay heat of the used fuel by about 40% at 1000 years
exit compared to LWR used fuel that is not reprocessed and trans-
muted in a CANDU reactor. This could have signicant impact on
the capacity of a geological repository.
Table 9
Input isotopic composition of the transuranic nuclides.
Isotope % by weight
Np-237 4.7
Pu-238 1.3
Pu-239 59.2
Pu-240 20.1
Pu-241 3.0
Pu-242 3.8
Am-241 9.9
Am-243 0.76
Cm-243 0.001
Cm-244 0.072
Cm-245 0.12
Cm-246 0.001
Full-core calculations of a TRU-MOX core have been completed
to demonstrate the feasibility of fuelling a CANDU 6 core with this
fuel. Several changes are needed with respect to the natural ura-
nium(NU) fuel cycle. The fuelling scheme for NUgenerally involves
replacing eight bundles in a fuel channel in each fuelling opera-
tion, referred to as an eight-bundle shift. Due to the higher initial
reactivity of the MOX bundle, in order not to introduce too much
localizedreactivity into the core, the fuelling scheme is reducedto a
combination of one- and two-bundle shifts, applied to 264 and 116
channels respectively. A detailed study on the impact on fuelling
machine utilization has not been completed, but the higher bur-
nup and therefore longer dwelling time of the fuel compensates
for the smaller number of bundles shifted during refuelling. This
results in a refuelling rate of about 3 bundles per day, compared to
15 bundles per day in the natural uranium CANDU 6 reactor.
A time-average model of the core was created to examine the
maximum channel and bundle powers. In addition to this model,
aninstantaneous snapshot model was also examined. The instanta-
neous model, which generates random ages for the fuel channels,
Table 10
The amounts of the transuranic nuclides that are transmuted in the TRU-MOX
scenario.
Nuclide % Transmuted Mass transmuted (kg/year)
Np-237 51.0 18.9
Total Np 48.1 17.9
Pu-238 206.6 21.2
Pu-239 78.6 348.4
Pu-240 2.4 3.8
Pu-241 68.8 16.5
Pu-242 128.2 38.4
Total Pu 41.1 276.8
Am-241 84.2 65.8
Am-243 221.1 13.3
Total Am 62.2 52.4
Cm-242 6.1
Cm-243 2774.4 0.22
Cm-244 1676.5 9.5
Cm-245 158.6 0.15
Cm-246 1706.0 0.13
Total Cm 2374.4 16.1
Total TRU 41.9 330.2
4802 B. Hyland, B. Gihm / Nuclear Engineering and Design 241 (2011) 47944802
Table 11
The maximum channel and bundle powers for the time-average and instantaneous
models for the MOX fuel cycle.
Time-average
model
Instantaneous
model
Maximum channel power (kW) 6300 7200
Maximum bundle power (kW) 820 910
was used to analyze the power increase that occurs when refu-
elling the reactor, referred to as the refuelling ripple. Values for the
maximum channel and bundle powers are given in Table 11. As is
expected, the maximumchannel and bundle powers are higher for
the instantaneous model, which accounts for an increase in power
as fresh fuel bundles are added to the core upon refuelling. These
values are all within normal CANDU reactor operating conditions.
3. Conclusion
In this paper, several different actinide transmutation schemes
in a CANDU reactor were presented. These methods are classied
under two categories: heterogeneous and homogeneous actinide
loading in CANDU reactors. In the former method, actinides are
conned to target pins or bundles and loaded in the reactor sepa-
rately from the rest of fuel. In the homogenous methods, actinides
are homogeneously mixed with fuel and placed in all channels in
the reactor.
The advantages of these methods are that they require mini-
mal operational and fuel changes in existing CANDU reactors. In
the case of using the centre pin of the bundle to carry the actinides,
the concept is very similar to the low void reactivity fuel (LVRF)
CANDU fuel design that contains a centre neutron absorber and
slightly enriched uraniumin the remaining fuel pins. In the hetero-
geneous channel method where actinide target bundles are placed
in dedicated transmutation channels of the reactor, the periph-
ery channels are selected to minimize the operational effect while
utilizingneutrons inthereector region. On-power refuellingcapa-
bility of a CANDU reactor allows the residence time of actinide
targets to differ signicantly from fuel bundles. With high neutron
uence, very high destruction of the actinides can be achieved. The
disadvantage of this scheme is that the total amount of actinide
destruction is limited since only small part of the reactor is being
used for actinide transmutation. The heterogeneous methods are
suitable when there are numbers of existing CANDU reactors that
can support a LWR eet.
Two different homogeneous actinide fuels are presented in this
paper corresponding to different actinide partitioning schemes:
separated Am mixed with LEU, and group-extracted actinides
mixed with natural uranium. Signicant actinide mass destruction
can be achieved using whole core loading of these homogeneous
actinidefuels. TheactinidetransmutationinasingleCANDUreactor
is in the order of several kilograms per year, achieving the support
ratio between 11 and 20 depending on the actinide-partitioning
scheme. The limitation in the homogeneous method is that the
fuel and targets are subject to the same neutron uence, lower-
ing the fraction of actinides destroyed during the irradiation in the
core. The discharge actinide fuel may require secondary reprocess-
ing and transmutation to achieve high destruction comparable to
the heterogeneous methods. A higher fuel burnup will increase the
actinide destruction fraction.
Several different strategies are being studied to maximize the
CANDU reactor utilization for actinide destruction. The currently
existing CANDU reactors can process small quantities of actinides
without signicant operational changes. This paper has not exam-
ined actinide destruction using thorium as a fuel matrix, rather
than uranium. Thorium is anticipated to be a better transmutation
matrixthanuraniumbecausefewer higher mass actinides wouldbe
producedduringirradiationandvaluableU-233wouldbeproduced
for subsequent recycle.
References
Altiparmakov, D., 2008. Newcapabilities of thelatticecodeWIMS-AECL. In: PHYSOR-
2008, International ConferenceonReactor Physics, Nuclear Power: ASustainable
Resource , Interlaken, Switzerland.
Boczar, P.G., et al., 1996. Advanced CANDU systems for plutonium destruction. In:
NATO Advanced Research Workshop on Advanced Nuclear Systems Consuming
Excess Plutonium , Moscow, Russia.
Chan, P.S.W., et al., 1997. CANDUa versatile reactor for plutonium disposition or
actinide 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 from
the 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 uraniumin CANDUreactors, in the US
GNEP program. The American Nuclear Society and the European Nuclear Soci-
ety 2007 International Conference on Making the Renaissance Real, November
1115, 2007, Washington, D.C. Trans. Am. Nucl. Soc. 97, 107108.
Forsberg, C.W., et al., 2004. Can thermal reactor recycle eliminate the need for
multiple repositories? In: 8th OECD/NEA Information Exchange Meeting on
Partitioning and Transmutation , Las Vegas.
Hyland, B., Dyck, G.R., 2007. Actinide burning in CANDU reactors. In: Global 07
Conference on Advanced Nuclear Fuel Cycles and Systems , Boise, Idaho.
Hyland, B., et al., 2008. Transmutation of actinides in CANDUreactors. In: OECD/NEA
Information 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 Options
and Industrial Perspectives , Paris, France.
Modolo, G., et al., 2008. Development and demonstration of a new SANEX process
for actinide(III)/lanthanide(III) separation using a mixture of CyMe4BTBP and
TODGAas selectiveextractant. In: Proceedings of the10thInformationExchange
Meeting 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 2009
Conference 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 standardtool computer programfor CANDU
reactor coredesignandanalysis. In: Proceedings of the13thPacic BasinNuclear
Conference , Shenzhen, China, 2002 October 2125.