Feasibility of a Breed-and-Burn Molten Salt Reactor

Full text

1174
Transactions of the American Nuclear Society, Vol. 116, San Francisco, California, June 11–15, 2017
Recent Advancements in Liquid and Solid Molten Salt Reactors—I
Feasibility of a Breed-and-Burn Molten Salt Reactor
Michael Martin, Manuele Aufiero, Ehud Greenspan, Massimiliano Fratoni
University of California, Berkeley, Department of Nuclear Engineering, Berkeley, CA 94720-1730 USA
michael.martin@berkeley.edu, manuele.aufiero@berkeley.edu, gehud@berkeley.edu, maxfratoni@berkeley.edu
INTRODUCTION
This work addresses the feasibility and characteristics of
the breed and burn (B&B) cycle for Molten Salt Reactors.
In the general B&B cycle, natural Uranium or Thorium is
used as fuel for a reactor which breeds a sufficient amount
to maintain criticality at equilibrium without reprocessing.
Interest in applying this scheme to an MSR has developed
recently [
1
] For the MSR B&B scheme considered here, fresh
salt containing fertile material is continually fed into the core
while salt with the composition of the homogeneous fuel is
expelled without reprocessing. Through an analysis of the
equilibrium conditions and startup requirements, the feasibility
of such a cycle operating with various salts is assessed.
MSR B&B EQUILIBRIUM
For a given salt, fertile material, and power level, the
equilibrium core composition will be determined by the salt
feed/removal time constant. To find this equilibrium compo-
sition, long burnup calculations were run using a modified
version of Serpent [
2
] that removes all isotopes from the fuel
with a given time constant. These dumped elements were re-
placed with an adequate amount of fresh fertile salt. The total
actinide molar share of the fuel was kept constant by altering
the proportion of fertile heavy metal in the feed stream. Each
fission product replaced one of the constituents of the origi-
nal salt to keep a constant total atomic density. In all cases,
noble metals and gasses were removed with a time constant
corresponding to an in-core halflife of 30min.
Infinite Geometry
To determine the feasibility of the MSR B&B cycle for
various salts, the equilibrium burnup calculations were con-
ducted using infinite media. The salts considered are listed in
table I. The densities of the salts, from top to bottom, were
taken from [
3
][
4
] and [
5
]. For each salt, equilibrium cal-
culations were run using both a natural Uranium feed and
pure Thorium feed employing various time constants. The
lithium in the lithium-fluoride salt was enriched to 99.99%
7Li
while the chlorine in the chloride salt was enriched to
99.99%
37Cl
. The power density of all salts was chosen to be
300
W/cm3
. The
kin f
for each salt as a function of burnup is
shown in figure 1. Neither of the cycles using fluoride salts
were feasible breeders because at equilibrium they present a
kin f
lower than unity for all time constants explored. How-
TABLE I: Salts considered for analysis; x =1 for all FP nuclides.
Salt Molar Proportions Density (g/cc)
(NaF+[FP]Fx)−KF −[Actinides]F443-24-33 4.263
(LiF+[FP]Fx)−[Actinides]F477.5-22.5 4.418
(NaCl+[FP]Clx)−[Actinides]Cl367-33 3.107
ever, the chloride salt showed breakeven potential employing
the Uranium-Plutonium cycle which presented peak excess
reactivity at equilibrium around a burnup of 0
.
39
FIMA
. This
burnup was achieved with a residence time (inverse of removal
time constant) around 9
year
. In terms of equilibrium
kin f
, the
Uranium-Plutonium cycle outperforms the Thorium-Uranium
cycle in all cases.
Fig. 1: Equilibrium
kin f
as a function of burnup. From top to bot-
tom:
(NaF+[FP]Fx)−KF −[Actinides]F4
,
(LiF+[FP]Fx)−[Actinides]F4
,
(NaCl+[FP]Clx)−[Actinides]Cl3.

1175
Transactions of the American Nuclear Society, Vol. 116, San Francisco, California, June 11–15, 2017 Transactions of the American Nuclear Society, Vol. 116, San Francisco, California, June 11–15, 2017
Recent Advancements in Liquid and Solid Molten Salt Reactors—I
Effect of Actinide Density
Several heavy metal chloride molar proportions were cho-
sen to investigate the effect of actinide density on the chloride
Uranium-Plutonium cycle equilibrium
kin f
. The results are
shown in figure 2. The increased actinide molar share in-
creases the excess reactivity at equilibrium for a given time
constant. The peaks of the curves shift to slightly higher resi-
dence times with increasing actinide density.
Effect of power level
The equilibrium
kin f
curve for the chloride Uranium-
Plutonium cycle at the arbitrarily chosen power density of
300
W/cm3
is compared with the same salt/cycle at a power
level of 100
W/cm3
in figure 3. The peak of the 100
W/cm3
curve occurs at a residence time approximately three times
longer than the 300
W/cm3
curve. The effect of changing the
power level is thus essentially a shift in the residence time
required to achieve the same
kin f
. The peak of the 100
W/cm3
curve is slightly higher than the 300
W/cm3
curve (1.10240
vs. 1.09910) which suggests better breeding performance at
lower power levels possibly due to lower insoluble and gaseous
fission product absorption.
Fig. 2: Effect of
[Actinides]Cl3
molar proportion. The equilibrium
kin f
curve
stretches upward with increasing actinide density.
Fig. 3: Effect of power level. A decrease in power density slightly improves
the equilibrium kin f for a given burnup level.
Finite Geometry
Equilibrium calculations were also run using finite re-
flected cores for the chloride salt with the Uranium-Plutonium
cycle at a power density of 300
W/cm3
and various heavy metal
densities. The core geometry was chosen to be a cylinder of
equal height and diameter. Both lead and steel reflectors were
employed. All reflectors were 1
m
thick on all sides, providing
an upper bound estimate. Each equilibrium calculation was
iteratively run with different core radii until a peak
kef f
within
300pcm of unity was achieved. This was chosen here to be the
minimum critical dimension. The minimum critical radii, ini-
tial Uranium content, initial Uranium enrichment, and burnup
for the core at equilibrium for each reflector and heavy metal
content combination is presented in table II.
TABLE II: Specifications for cores critical at equilibrium employing various
reflectors and heavy metal densities. A 1:1 diameter to height ratio was used.
[Actinide]Cl_3% Radius
(m)
Initial load
(MTU)
Initial Enrichment
(wt%)
Burnup
(FIMA)
Lead
33 2.30 240 11.20 0.403
40 1.95 157 11.20 0.397
50 1.70 112 11.20 0.407
Steel
33 2.80 432 11.50 0.404
40 2.45 312 11.58 0.409
50 2.25 258 11.47 0.432
STARTUP FEASIBILITY
The start-up of a chloride salt B&B reactor employing the
Uranium-Plutonium cycle as described here will require the
initial feed salt to contain fissile material. Choosing enriched
Uranium as the feed material keeps the MSR B&B cycle
reprocessing-facility free. To investigate the feasibility of such
a startup process, the same modified Serpent was employed
as for the equilibrium calculations. The lower bound of the
feed enrichment was chosen to be natural Uranium. The upper
bound on the enrichment was chosen to be that of the initial
core load.
The time constant and geometry for the minimum criti-
cal dimension core with 33% [
Actinide
]
Cl3
for both types of
reflectors was used. The results are shown in figure 4. There
is no reactivity drop in the cores fed with salt of the initial
enrichment level. In the steel reflected core fed with natural
Uranium, sufficient breeding has occurred at 2.08 years such
that all reactivity has been recovered. The lead reflected core
recovered from the initial reactivity drop in only 1.69 years.
These results indicate that the enrichment levels required for
the startup feed are manageable (no greater than the initial
requirement) and not long lived.
ENRICHMENT SAVINGS
The lifetime enrichment requirements of the MSR B&B
cycle were compared with those of an AP1000 [
6
]. The MSR
case taken as reference was the 33% heavy metal content, steel
reflected core operating at 300
W/cm3
. The power normalized
SWU requirements as a function of lifetime length are shown
in figure 5. The cumulative normalized separative work re-
quired for the operation of the AP1000 remains constant at 539

1176
Transactions of the American Nuclear Society, Vol. 116, San Francisco, California, June 11–15, 2017
Recent Advancements in Liquid and Solid Molten Salt Reactors—I
SWU/GWt ∗yr
while that for the MSR begins high around
9
.
01
∗
10
5SWU/GWt ∗yr
and decrease rapidly. The MSR was
assumed to have been fed using the initial enrichment level
(11.5%) for the amount of time it took the natural Uranium
fed core (figure 4) to recover its reactivity (2.08yrs). The inter-
section of the curves occurs at 5.91 years. This represents the
minimum amount of time the MSR must be run before SWU
savings occur.
Fig. 4: Startup
kef f
curves for steel (top) and lead (bottom) reflected cores
with natural and initial enrichment Uranium feeds.
Fig. 5: Comparison of the cumulative separative work required per unit energy
by the MSR B&B cycle and AP1000. The MSR presents SWU savings after
5.91 years.
CONCLUSIONS
The MSR B&B scheme studied here was shown to be
feasible when chloride salts and natural Uranium fuel were
employed. The equilibrium conditions showed adequate breed-
ing performance that resulted in cores of reasonable critical
dimensions with start-up feed requirements within practical
bounds. Future studies will address the feasibility of attaining
passive safety and of practically handling the radiation damage
to structures adjacent to the core. The waste characteristics,
proliferation resistance, and optimal method of approaching
equilibrium for the MSR B&B cycle are also to be addressed.
REFERENCES
1.
J. LATKOWSKI, “Emerging Commercial Design Con-
cepts: TerraPower,” in “Presented at the Workshop on
Molten Salt Reactor Technologiesâ˘
Aˇ
TCommemorating the
50th Anniversary of the Startup of the MSRE: From the
MSRE to a New Emerging Class of Reactors 50 Years
Later.”, Oak Ridge, TN. (October 14-15 2015).
2.
M. AUFIERO, A. CAMMI, C. FIORINA, J. LEPPÄNEN,
L. LUZZI, and M. RICOTTI, “An extended version of
the SERPENT-2 code to investigate fuel burn-up and core
material evolution of the Molten Salt Fast Reactor,” Journal
of Nuclear Materials,441,1, 473–486 (2013).
3. G. J. JANZ, Molten salts handbook, Elsevier (2013).
4.
O. BENES and R. KONINGS, “Molten salt reactor fuel
and coolant,” Comprehensive Nuclear Materials,
3
, 374
(2012).
5.
V. DESYATNIK, S. KATYSHEV, S. RASPOPIN, and Y. F.
CHERVINSKII, “Density, surface tension, and viscosity of
uranium trichloride-sodium chloride melts,” Soviet Atomic
Energy,39,1, 650 (1975).
6.
NRC, “Final Safety Evaluation Report Related to Certifica-
tion of the AP1000 Standard Design (NUREG-1793, Initial
Report),” (2004).