Note: Descriptions are shown in the official language in which they were submitted.
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Molten Salt Reactor
TECHNICAL FIELD
This disclosure relates to nuclear reactors, and more particularly to molten
salt
reactors.
BACKGROUND
Thermal-spectrum molten salt reactors have long interested the nuclear
engineering community because of their many safety benefits ¨ passive shutdown
ability,
low pressure piping, negative void and temperature coefficients, and
chemically stable
coolants ¨ as well as their scalability to a wide range of power outputs. They
were
originally developed at the Oak Ridge National Laboratory (ORNL) in the 1950s,
1960s,
and 1970s, and working versions were shown to operate as designed [1].
The bulk of the early work on these designs focused on component lifetime ¨
specifically, developing alloys able to maintain their mechanical and material
integrity in
a corrosive, radioactive salt environment. Experimental tests running over
several years
at ORNL in the 1960s and 1970s showed that modified Hastelloy-N possesses the
necessary chemical and radiation stability for long-term use in molten salt
reactors.
Despite this progress, the USA remained focused on light-water reactors for
commercial
use, primarily due to extensive previous experience with naval water-cooled
reactors.
Advocates of thorium and increasing demand for small modular reactors drove
renewed
examination of molten salt in the 1990s. In 2002, the multinational Generation
IV
International Forum (GIF) reviewed approximately one hundred of the latest
reactor
concepts and selected molten salt reactors as one of the six advanced reactor
types most
likely to shape the future of nuclear energy "due to advances in
sustainability, economics,
safety, reliability and proliferation-resistance" [2].
SUMMARY
An advanced molten salt reactor that generates clean, passively safe,
proliferation-
resistant, and low-cost nuclear power. This reactor can consume the spent
nuclear fuel
(SNF) generated by commercial light water reactors or use freshly mined
uranium at
1
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
enrichment levels as low as 1.8% U-235. It achieves actinide burnups as high
as 96%,
and can generate up to 75 times more electricity per ton of mined uranium than
a light-
water reactor.
Key characteristics of a first commercial plant are as follows:
Reactor Type Molten Salt Fueled Reactor
Fuel Uranium or spent nuclear fuel (SNF)
Salt LiF-(Heavy Metal)F4
Moderator Zirconium Hydride
Neutron Spectrum Thermal
Thermal Capacity 1250 MWth
Gross Electric Capacity 550 MWe
Net Electric Capacity 520 MWe
Outlet Temperature 650 C
Gross Thermal
44% using steam cycle with reheat
Efficiency
Fuel Efficiency 75X higher per MW than LWR
Long-lived Actinide
Up to 96% less per MW than LWR
Waste
Station Blackout Safety Walkaway safe without outside intervention
Overnight Cost $2 billion
Typically for base load;
Mode of Operation
May be used for load following
Transatomic Power has greatly improved the molten salt concept, while
retaining
its significant safety benefits. The main technical change we make is to
combine a
moderator and fuel salt that have not previously been used together in molten
salt
reactors: a zirconium hydride moderator with a LiF-(Heavy metal)F4 fuel salt.
Together,
these components generate a neutron spectrum that allows the reactor to run
using fresh
uranium fuel with enrichment levels as low as 1.8% U-235, or using the entire
actinide
component of spent nuclear fuel (SNF). Previous molten salt reactors such as
the ORNL
Molten Salt Reactor Experiment (MSRE) relied on high-enriched uranium, with
33% U-
235 [1]. Enrichments this high are no longer permitted in commercial nuclear
power
plants.
Transatomic Power's design also enables extremely high burnups ¨ up to 96% ¨
over long time periods. The reactor can therefore run for decades and slowly
consume the
actinide waste in its initial fuel load. Furthermore, our neutron spectrum
remains
primarily in the thermal range used by existing commercial reactors. We
therefore avoid
2
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
the more severe radiation damage effects faced by fast reactors, as thermal
neutrons do
comparatively less damage to structural materials.
Some radioactive materials release neutrons. When a neutron strikes a fissile
atom, such as U-235, at the right speed, the atom can undergo "fission" or
break into
smaller pieces, which are called fission products, and produce free neutrons.
Fission
breaks bonds among the protons and neutrons in the nucleus, and therefore
releases vast
amounts of energy from a relatively small amount of fuel. Much of this energy
is in the
form of heat, which can then be converted into electricity or used directly as
process heat.
Most neutrons travel too quickly to cause fission. In a typical nuclear
reactor, the
fuel is placed near a moderator. When neutrons hit the moderator they slow
down, which
makes them more likely to cause fission in uranium. If the average number of
free
neutrons remains constant over time, the process is self-sustaining and the
reactor is said
to be critical.
Despite the use of the word critical, there is no chance of an atomic
explosion in
nuclear power plants. The fuel used in civilian nuclear reactors has a low
enrichment
level that is simply not capable of achieving the chain reaction required for
an atomic
explosion. The main concern in nuclear power is to avoid a steam explosion,
fire, or
containment breach that could allow the release of radioactive materials
outside the plant
and affect public health.
Light-water nuclear reactors ¨ the most prevalent kind of reactor in use today
¨
are fueled by rods filled with solid uranium oxide pellets. The fuel rods are
submerged in
water. Water is a moderator that slows neutrons to the correct speed to induce
fission in
the uranium, thereby heating up the rods. The water also carries heat away
from the rods
and into a steam turbine system to produce electricity. A key problem with
water is risk
of steam explosion if the reactor's pressure boundary or cooling fails.
In a molten salt reactor, a radioactive fuel such as uranium or thorium is
dissolved
into fluoride or chloride salts to form a solution that we call a "fuel salt."
The fuel salt is
normally an immobile solid material, but when heated above approximately 500
C, it
becomes a liquid that flows. Thus it is the liquid fuel salt, rather than
water, that carries
the heat out of the reactor. The plant can operate near atmospheric pressure
with a coolant
3
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
that returns to a solid form at ambient temperatures. This feature simplifies
the plant and
assures greater safety for the public.
Molten salt reactors are quite different from sodium fast reactors, even
though
many people think of sodium when they hear of salt. The sodium metals used by
those
reactors can release a hydrogen byproduct that is combustible in the presence
of air or
water. Our fluoride salts remove this fire risk, while further simplifying and
increasing
the safety of the plant design.
A version of our reactor can also operate using thorium fuel. Thorium has
special
merit as a nuclear fuel due to its generally shorter-lived waste and higher
potential burn-
up. The TAP reactor can also achieve the same benefits from uranium, which has
an
existing industrial base. Using uranium also lets us create a reactor that can
slowly
consume the world's existing stockpiles of spent nuclear fuel and,
potentially, stockpiles
of plutonium as well, thereby providing a great benefit to society.
The details of one or more embodiments of the invention are set forth in the
accompanying drawings and the description below. Other features, objects, and
advantages will be apparent from the description and drawings, and from the
claims.
DESCRIPTION OF DRAWINGS
Figure 1 is a schematic of the TAP reactor, showing the reactor vessel,
primary
loop, intermediate loop, and drain tanks.
Figure 2 is a simplified reactor schematic, showing the primary loop,
intermediate
loop, drain tank, and outlet to the fission gas processing system.
Figure 3 is a temperature profile of a light water reactor's solid fuel pin,
from
center to edge.
Figure 4 shows decay heat density in an LWR and a TAP reactor.
Figure 5 is a cooling curve for fuel salt in auxiliary tank with 25 MW of
cooling.
Figure 6 compares temperature progression effects for a light water reactor
(LWR) and a TAP reactor.
Figure 7 compares the neutron spectrum in a zirconium hydride moderated TAP
reactor, a graphite moderated molten salt reactor, and a fast spectrum molten
salt reactor.
4
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Figure 8 compares electricity production per metric ton of natural uranium in
a
light water reactor and a TAP reactor.
Figure 9 compares mass percentages of important actinides as a function of
time
in a TAP reactor.
Figure 10 plots the multiplication factor of an infinite lattice of varying
moderator
and fuel-salt volume fractions.
Figure 11 shows the effect of enrichment (fissile concentration) on burnup as
a
function of conversion ratio.
Figure 12 plots conversion ratio as a function of fuel-salt volume fraction.
Figure 13 is a schematic of a two-region reactor core.
Figure 14 is a schematic of a two-region core with central unmoderated region.
Figure 15 is a schematic of a three region core with two distinct ratios of
fuel-salt
to moderator volumes.
Figure 16 is a schematic of a three region core with three distinct ratios of
fuel-
salt to moderator volumes.
Like reference symbols in the various drawings indicate like elements.
DETAILED DESCRIPTION
Reactor Description and Design Considerations
We begin by describing the components of the TAP reactor that are within and
adjacent to the nuclear island and discuss design considerations. We show a
rendering
and schematic of the nuclear island, describe the benefits of liquid fuel as
compared to
solid fuel, and then review the zirconium hydride moderator, corrosion,
reactor
neutronics, and waste stream.
Nuclear Island Rendering and Schematic
Figure 1 shows a rendering of the TAP reactor seated in a concrete nuclear
island
structure for a 520 MWe nuclear power plant incorporating a TAP reactor. This
same
system is shown schematically in Figure 2.
5
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
The reactor's primary loop contains the reactor vessel (including the
zirconium
hydride moderator), pumps, and primary heat exchanger. Pumps continuously
circulate
the LiF-(Heavy metal)F4 fuel salt through the primary loop. The pumps,
vessels, tanks,
and piping are made of modified Hastelloy-N, which is highly resistant to
radiation and
corrosion in molten salt environments. Within the reactor vessel, in close
proximity to the
zirconium hydride moderator, the fuel salt is in a critical configuration and
steadily
generates heat.
The heat generated in the primary loop is transferred via heat exchangers into
intermediate loops filled with molten LiF-KF-Na-F (FLiNaK) salt, which does
not
contain radioactive materials. The intermediate loops in turn transfer heat to
the steam
generators. The intermediate loops therefore physically separate the nuclear
material
from the steam systems, adding an extra layer of protection against
radioactive release.
The steam generators use the heat from the intermediate loop to boil water
into
steam, which is then fed into a separate building that houses the turbine. The
reactor runs
at a higher temperature than conventional reactors¨the salt exiting the
reactor core is
approximately 650 C,whereas the core exit temperature for water in a light
water reactor
is only about 330 C (for a pressurized water reactor) or 290 C (for a boiling
water
reactor). The thermal efficiency when connected to a standard steam cycle is
44%, as
compared to 34% in a typical light-water reactor. The higher efficiency
directly reduces
cost because it permits smaller turbines ¨ turbines are a major expense for
nuclear power
plants.
The nuclear island also contains fission product removal systems. The majority
of
fission product poisons are continuously removed via an off-gas system (not
shown in
Figure 1). As these byproducts are gradually removed, a small amount of fuel
(either SNF
or low-enriched fresh fuel) is regularly added to the primary loop. This
process maintains
a constant fuel mass, and allows the reactor to remain critical for decades.
Through
continuous fueling and filtering of key fission product poisons we are able to
process the
initial fuel load in the reactor for long periods of time, on the order of
decades, as
compared to a typical 4 year lifetime in a light water reactor. During this
time, nearly all
of the actinide fuel is converted into fission products and energy.
6
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Liquid Fuel vs. Solid Fuel
Nearly all currently operating commercial reactors use solid uranium oxide as
fuel. The uranium oxide, which is in the form of solid pellets, is surrounded
by a metal
cladding that helps the fuel retain its shape within the reactor. In contrast,
Transatomic
Power's reactor uses liquid fuel instead of solid fuel pins. We dissolve
uranium (or SNF)
in a molten fluoride salt, which acts as both fuel and coolant.
Liquid fuel offers significant advantages during normal operation. Primarily,
it
allows for higher reactor outlet temperatures, which lead to higher overall
thermal
efficiency for the plant.
Higher Outlet Temperatures
In a commercial light water reactor, water is used as a working fluid to carry
the
heat away from the hot outer surface of the fuel cladding, typically at about
330 C, to the
plant's power loop. A higher cladding temperature allows for a higher water
temperature,
which allows for a more efficient power production cycle. A problem with solid
fueled
reactors, however, is that the uranium oxide material is a poor heat
conductor. As shown
in Figure 3, the centerline temperature of the fuel pin must be very high ¨ up
to 2000 C
in a pressurized water reactor (PWR) ¨ to generate an acceptably high
temperature on the
outer wall of the cladding. In most light water reactors, it is not possible
to increase the
outer cladding temperature significantly beyond 330 C, because that would
result in an
unacceptably high fuel centerline temperature.
A liquid-fueled reactor does not have these problems, because the fuel and
coolant are the same material. The fuel salt is a good heat conductor, and
therefore can
have both a lower peak temperature and a higher outlet temperature than a
solid fueled
reactor.
Decay Heat is Better Distributed
One major safety advantage of liquid fuel is that it is significantly easier
to cool it
down during an accident scenario, as compared to solid fuel. Adequately
cooling the fuel
is crucial during an accident, because the fuel continues to produce decay
heat even after
the system becomes subcritical.
7
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
The fuel in Transatomic Power's reactor is dissolved and diluted across a
substantial mass of salt, which distributes the decay heat and allows for
easier cooling
than an equivalently-sized solid fueled reactor. Figure 4 compares the decay
heat density
(MWth of decay heat per cubic meter of fuel) in a TAP reactor and an LWR over
time.
The TAP reactor's lower decay heat density makes it easier to contain and cool
the liquid fuel during an accident.
Easier to Remove Decay Heat
Solid fueled reactors must bring coolant to their fuel in an accident
scenario. If
either coolant or cooling power is lost, decay heat production can quickly
raise the
reactor core temperature to levels high enough to severely damage its
structure.
Light-water reactors were originally invented for use in submarines, which can
use the ocean as an effectively infinite heat sink. On land, commercial power
plants must
reserve enough water in tanks and enough battery power in pumps to sustain
emergency
cooling for approximately a day, until help can arrive with more water and
power. The
most advanced plants now being built in the US will be able to extend the self-
sufficiency
period to 72 hours. However, local aid may or may not be available by then. As
recent
events at Fukushima demonstrated, a breakdown in transportation infrastructure
to
deliver emergency assistance can greatly exacerbate a reactor accident.
Unlike solid fueled reactors, liquid fueled reactors can drain fuel directly
out of
the core. This drainage can happen quickly, without pumping, through the use
of passive
safety valves and the force of gravity. One such passively safe drainage
mechanism,
called the freeze valve, was tested repeatedly with success during the ORNL
MSRE [1].
A freeze valve consists of a drain in the reactor leading to a pipe that is
plugged by a
solid core of salt. The salt remains solid via electric cooling. If the
reactor loses external
electric power, the cooling stops, the plug melts, and fluoride salt drains
out of the reactor
core into an auxiliary containment vessel. Fission ceases because the fuel is
separated
from the moderator and because of the relatively high surface area geometry of
the
auxiliary tank. The high surface area to volume ratio in the auxiliary tank
allows molten
salt reactors to effectively change their fuel geometry to speed cooling after
an accident.
8
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
The decay heat of the auxiliary tank is low enough to be removed by natural
convection via a cooling stack, thereby eliminating the need for electrically-
pumped
coolant. A NaK cooling loop in the auxiliary tank is connected to a stack and
allows for
25 MW of passive cooling to the fuel, adequate to air-cool the entire fuel
salt inventory
from liquid to solid state within 1.5 to 3 hours without outside power or
coolant. Figure 5
shows the temperature of the fuel salt inventory in the auxiliary tank as a
function of time
with 25 MW of cooling. The upper and lower bounds for the cooling curve are
shown as
dashed lines. Thermal data for the salt is based on molecular dynamics
simulations [3]
and extrapolated experimental data [4].
Slower and Less Catastrophic Accident Progression
Figure 6 shows the different consequences of unchecked fuel heating in an LWR
and a TAP reactor. As shown in the "LWR" column of Figure 6, partial cooling
is helpful
but not sufficient in an accident scenario. Even after the reactor becomes
subcritical, the
fuel pins continue to generate heat from delayed neutron interactions.
The risk of a steam flash or rupture and release exists during accidents at
any
temperature above 100 C, the boiling point of water at atmospheric pressure.
Starting at
approximately 700 C, Zircaloy and water together generate significant amounts
of
hydrogen. The reaction becomes exothermic above 1200 C, as the reaction
produces heat
more quickly than it can be removed ¨ this further raises temperatures and
runs counter to
cooling efforts. The hydrogen generation can lead to a fire or explosion (as
happened at
Fukushima), and damage to the cladding releases radioactive materials that
could travel
away from the plant if they escape containment. Steam and fire are driving
forces that
increase the distance such materials could travel.
After an emergency, these overheating accident scenarios can develop within a
few hours. A light-water reactor core, filled with solid fuel pins that are
poor heat
conductors, requires a cooling period of months or years to reach a stable
cladding
temperature of 100 C or below. This mismatched timing ¨ hours to overheat
versus many
months to cool off ¨ is what makes nuclear safety for light-water reactors
enormously
challenging, and leaves these reactors particularly vulnerable to disasters
that were not
anticipated at the design stage, known as "beyond design basis" accidents.
9
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Molten salt reactors avoid these issues inherently ¨ by their choice of
materials.
As shown in the "Transatomic Power" column in Figure 6, a molten salt reactor
operates
at a peak temperature of 650-700 C, far below the salt's boiling point of
approximately
1200 C. The reactor's steady-state operation is already in the "green" zone.
The thermal
mass of the fuel is now an asset instead of a challenge, because it serves to
resist any
sudden heat increase. If the reactor temperature were to climb, temperatures
greater than
700 C passively melt a freeze valve (discussed in the "Better Inherent Safety"
section of
this paper), which drains fuel from the reactor and allows it to flow into a
subcritical
configuration with a high surface area. The subcritical molten salt still
generates decay
heat, but the high surface area allows it to readily cool down via natural
convection and
conduction.
At the other end of the temperature spectrum, the salt safely freezes in place
if
temperatures drop below 500 C. Unlike water, the salt becomes denser after it
freezes, so
this condition does not increase system pressure. As the TAP reactor operates
at
atmospheric pressure and has few conditions that could create strong driving
forces, the
solid salt is likely to remain safely in containment and within the exclusion
zone of the
plant.
In addition to the inherent safety benefits of molten salt liquid fuel, the
TAP plant
design has additional safety features and containment strategies for defense
in depth.
These safety features and strategies are discussed further below.
Salt Formulation
The vast majority of past work on molten salt reactors has used a lithium-
beryllium-fluoride salt, called FLiBe. Transatomic Power's reactor instead
uses LiF-
(Heavy metal)F4 fuel salt. One known drawback of this salt is that its melting
point is
higher than that of FLiBe, and thus the primary loop piping must be carefully
designed to
avoid cold spots that could restrict flow and induce freezing in the salt. We
chose to
accept this engineering challenge for two reasons.
The first reason is that FLiBe contains beryllium. A small fraction of the
population is hypersensitive to this material, and even trace amounts of
beryllium can
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
induce the chronic lung disease berylliosis in these people. We therefore
choose a fuel
salt that does not contain beryllium.
The second reason is that LiF-(Heavy metal)F4 is capable of containing a
higher
concentration of uranium than FLiBe salt. Therefore, each liter of our fuel
salt has a
higher amount of uranium than would be possible using FLiBe. This salt
composition
thus helps us operate using low-enriched fuels, as well as spent nuclear fuel.
Zirconium Hydride Moderator
A key difference between Transatomic Power's reactor and other molten salt
reactors is its zirconium hydride moderator, which we use instead of a
conventional
graphite moderator. The reactor's critical region contains zirconium hydride
rods. These
rods are surrounded by cladding to extend the life of the moderator in the
corrosive
molten salt.
The available experimental data suggest that the service lifetime of the
moderator
rods will be at least 4 years. Additional in situ testing is needed to
determine how far that
lifetime can be extended. Ultimately, it may not be necessary to replace the
zirconium
hydride moderator assemblies over the lifetime of the plant. Our first design
provides for
maintenance access to the rods for evaluation and replacement, although this
feature may
be eliminated in a future version.
Using this moderator is an important advancement. Early molten salt reactors,
such as the MSRE, used a graphite moderator that would shrink and swell over
time
under irradiation [1]. These dimensional changes not only reduced mechanical
integrity,
they also complicated reactor operation, since the degree of change and
quality of
moderation varied over time and spatially within the core. This variability
made it
necessary to replace the graphite every 4 years. In contrast, zirconium
hydride moderator
rods experience substantially less volumetric change than graphite under
neutron
irradiation [5].
In the design for the ORNL Molten Salt Breeder Reactor, 80-90% of the core
volume was occupied by the graphite, leaving only 10% - 20% of the core for
fuel salt. It
was therefore necessary to enrich the uranium in the fuel salt to 33% U-235
[1]. This high
11
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
enrichment level was acceptable for a US national lab experiment; however, it
is above
modern limits of 20% U-235 for research reactors and well above the 3-5% U-235
enrichment level that is typical of commercial power reactors. Higher
enrichments are
discouraged as a proliferation concern.
By comparison, zirconium hydride's high hydrogen density allows it to achieve
the same amount of thermalization as graphite in a much smaller volume. The
zirconium
hydride moderator therefore allows us to significantly reduce the reactor core
volume,
thereby reducing the size and cost of the reactor vessel and the volume of
fuel salt. In
Transatomic Power's reactor, only about 50% of the core volume is moderator,
which
gives us room for five times more fuel salt in the same size core, allowing
better
performance, reduced enrichment, and lower cost.
Co-optimizing the core geometry with the new moderator and new salt
formulation, we can drop the minimum fuel enrichment level from 33% to 1.8%.
This
efficiency also enables us to consume SNF.
One of the factors we examined in selecting a zirconium hydride moderator is
the
stability of hydrogen in zirconium hydride at high temperature and under
irradiation. The
available data are extensive, and show that zirconium hydride is stable at the
temperatures and neutron fluxes present in Transatomic Power's reactor [6-10].
The
Soviet TOPAZ reactors, which generated thermionic power for satellites,
demonstrated
the effectiveness of their zirconium hydride moderator in experimental tests
on the
ground and in orbit [11]. According to experimental tests performed in
conjunction with
the TRIGA [6] and SNAP [7] reactors, both of which used uranium zirconium
hydride
fuel, zirconium hydride remains stable in a reactor core at temperatures at
least up to
750 C. According to Simnad, "... zirconium hydride can be used at temperatures
as high
as 750 C under steady-state and 1200 C under short transient pulse operation"
[6].
Modest hydrogen redistribution may occur within the moderator, because there
exists a temperature gradient within the moderator rod. The moderator is
internally
heated through gamma heating and neutron scattering, and the centerline
temperature of
the moderator rod will therefore be approximately 50 C higher than the wall
temperature.
Some experimental data are available for temperature gradient-driven hydrogen
diffusion
12
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
in zirconium hydride. Huangs et al. tested a temperature gradient of 140 C in
a ZrH1.6
rod, with a centerline temperature of 645 C and a surface temperature of 505 C
[8]. Their
steady-state result showed ZrH1.7 on the surface and ZrH1.5 at the centerline
[8]. Our
research indicates that this hydrogen concentration gradient, or even a
gradient several
times larger than this, would not be detrimental to reactor function.
Additional work by Ponomarev-Stepnoi et al., in which zirconium hydride blocks
were thermally cycled up to 650 C, found "statistically negligible" hydrogen
emission
after 4.1 years, and a maximum of 2% emission after 10 years of thermal
cycling [9].
We conclude that significant hydrogen outgassing will not occur in this
reactor
under normal operation. If significant hydrogen outgassing does occur through
some
unknown condition, the zirconium hydride moderator becomes less effective
(because of
the lower amount of hydrogen present), and thereby reduces reactivity in the
core.
Zirconium on its own essentially does not moderate neutrons. Free hydrogen
diffuses
through the cladding and into the salt, where it bubbles out and is removed
continuously
by the outgas system. This feature bears some similarity to the inherent
safety of
uranium-hydrogen fuel used in TRIGA reactors, and represents an added safety
benefit
over previous molten salt reactors. Even in an extreme accident scenario,
including
failure of the off-gas removal, the system is designed so that the hydrogen
concentration
is never high enough to lead to a hydrogen explosion.
Corrosion
The reactor's primary loop piping, reactor vessel, valves, pumps, and heat
exchangers are made with modified Hastelloy-N. This alloy is corrosion-
tolerant in
molten salt environments.
Hastelloy-N and modified Hastelloy-N were developed specifically for molten
fluoride systems, and have generally good corrosion resistance in molten
fluoride salt
environments [12]. The Molten Salt Breeder Reactor (MSBR) project at the Oak
Ridge
National Laboratory concluded that modified Hastelloy-N is a suitable material
for
molten salt reactors from a corrosion standpoint [12]. Furthermore, MSBR
research
concluded that modified Hastelloy-N suffers much less radiation embrittlement
than
13
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
unmodified Hastelloy-N, the previous formulation of the alloy used in the MSRE
[12].
Aside from the reduced radiation embrittlement, the material properties of
modified
Hastelloy-N are, according to MSBR research, "generally better" than those of
Hastelloy-
N [12].
There are some additional concerns related to the mechanical integrity of the
primary loop piping. The first is the possibility of mechanical fatigue and
subsequent
crack initiation due to thermal striping, in which temperature fluctuation
occur at the
interface between two fluid jets at different temperatures. Fluid dynamics
simulations of
the reactor vessel can partially predict these effects, and they will be
further tested via
experiment in the early stages of the work.
The second concern relates to welding and joining issues in the primary loop.
The
piping joints are the weakest links in the primary loop, and it is important
to make sure
that they retain their mechanical and material integrity throughout reactor
operation.
Furthermore, it is important to ensure that the metal used in brazing or other
joining
techniques is compatible with the molten salt, and doesn't exacerbate
corrosion effects.
Prior research shows that nickel-based brazing alloys are compatible with high-
temperature molten salts [13].
One benefit is that the molten salt reactor piping and vessel walls are
thinner than
those of a light water reactor (because of the lower-pressure piping in a
molten salt
reactor), which reduces the possibility of inadvertently stressing the metal
while welding.
Welding and joining issues will be tested experimentally in small-scale test
loops.
In the future, the reactor may be adapted to use high-temperature ceramics,
such
as SiC-SiC fiber composites, in place of Hastelloy components. These ceramics
are not
yet being manufactured on an industrial scale, but will likely be available
within 5 to 10
years. Moving from metals to ceramics will allow us to further increase the
reactor's
operating temperature, thereby increasing the system's thermal efficiency and
enabling a
broader range of process heat applications.
14
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Neutronics, Fuel Capacity, and Waste Stream
Reactor Neutronics
Molten salt reactors are versatile in terms of fuel: they can be powered by a
range
of different fissionable materials, including uranium, plutonium, and thorium.
Although
Transatomic Power's approach could potentially be used with thorium, we are
initially
focused on the uranium-plutonium cycle. This fuel cycle allows us to power the
reactor
with either uranium from an existing industry supply chain or, ideally, to use
a fleet of
TAP reactors to consume and substantially eliminate the nation's stockpiles of
SNF.
Conventional wisdom holds that only a fast reactor can effectively burn SNF.
This statement, however, assumes a system in which solid nuclear fuel must be
regularly
replaced due to the build-up of fission product gases and radiation damage.
Under these
assumptions, only fast reactors have neutron economies that can destroy enough
actinides
during a fairly short window of time. In a fast reactor, this actinide burning
is
accomplished by keeping neutrons at high kinetic energies, where the fission-
to-capture
ratio is high, with the drawback that the reactor core is exposed to extremely
challenging
radiation damage.
There are other ways of achieving a neutron spectrum capable of burning SNF.
For example, thermal-spectrum CANDU reactors are able to run on spent nuclear
fuel by
using on-line refueling and a more efficient moderator (heavy water instead of
light
water) to reduce neutron capture. However, burnup in CANDUs is also limited by
the
accumulation of fission product poisons that are trapped in the fuel rods. The
TAP reactor
circumvents this limitation by continuously removing fission products from its
liquid
fuel.
As described previously, the Transatomic Power reactor burns the same fuel for
decades. The combination of the TAP reactor's particularly efficient neutron
economy,
which allows it to run on fuel with very low enrichment levels, and molten
salt reactors'
general ability to continuously remove fission products from the fuel are what
together
enable us to destroy SNF. More generally, they allow us to achieve high
efficiency for a
clean and complete burn with very little waste.
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Figure 7 compares the neutron energy spectra in an unmoderated molten salt
reactor, one moderated with ZrH1.6, and one moderated with graphite. The
reactor
moderated with ZrH1.6 has significantly more neutrons in the thermal region,
defined as
neutrons with energies less than approximately 1 eV, thereby allowing it to
generate
power from low-enriched uranium or spent fuel using the U-Pu fuel cycle. The
epithermal (approximately 1 eV ¨ 1 MeV) spectrum is lower than that of
graphite, but
still sufficient to contribute to waste burning. The fast spectrum (greater
than 1 MeV) for
the zirconium hydride moderated reactor is greater than that of the graphite
moderated
reactor, and therefore contributes strongly to waste burning.
Fuel Capacity and World Uranium Reserves
When running on fresh fuel, the TAP reactor is able to generate up to about 75
times more electricity than a light water reactor per kilogram of natural
uranium ore, as
shown in Figure 8.
There are three factors driving this higher electricity output: lower
enrichment,
higher burn-up, and better conversion of heat to electricity:
Lower Enrichment: One ton of natural uranium ore yields 88 kilograms of LWR
fuel enriched to 5%. However, it yields 274 kilograms if only enriched to
1.8%. This is a
factor of 3.1X more starting fuel mass for the TAP reactor.
Higher Burn-up: At 5% enrichment, lightwater reactors have improved their
burnups from from 30 Gigawatt-days per metric ton of heavy metal (GWd per
MTHM),
and are quickly approaching burnups as high as 45 GWd per MTHM. In contrast,
the
TAP reactor can achieve up to 96% burnup at 1.8% enrichment ¨the equivalent of
870
GWd per MHTM out of a theoretical maximum of 909 GWd per MHTM. This is a
factor
of 19.2X more thermal energy for the TAP reactor.
Better Conversion: Light water reactors have outlet temperatures of 290 C -
330 C, and typical thermal efficiencies of about 34%. TAP reactors have an
outlet
temperature over 650 C with a gross thermal efficiency of about 44%. This is a
factor of
1.3X more for the TAP reactor.
Proven world reserves of uranium are estimated to be about 6 million metric
tons
if the market price were $250 per kilogram (current prices are about $130 per
kilogram ¨
16
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
at a higher price more mines are viable). Using light-water reactors, these
reserves are
only enough for about three million terawatt-hours of electricity. However,
the world
consumes about 20,000 terawatt-hours of electricity annually, and this rate is
set to triple
by 2030 as we climb toward a steady global population of ten billion people.
LWRs can
therefore only fully supply world electricity needs for about 50 years, even
at twice
today's uranium prices.
This limitation is currently not an alarming problem because, at this point,
nuclear
power provides only 12% of global electricity generation ¨ there are several
centuries of
uranium available at this current generation rate. If, however, nuclear
power's generation
share increases as countries turn away from fossil fuels, LWRs comparatively
low
burnups may become an issue. By comparison, the TAP reactor can use current
known
uranium reserves to supply 100% of the world's electricity needs for 3,500
years.
Techniques now under research around the world for collecting uranium from
seawater are estimated to become economically viable once uranium reaches a
price of
about $300 per kilogram. The TAP reactor generates enough electricity per
kilogram of
fuel that it remains commercially viable even with extremely high uranium
prices. The
TAP reactor can therefore enable a greater degree of energy dependence for
nations
without significant domestic uranium production, such as France, Japan, South
Korea,
UK, Spain, Argentina, and India. (Key uranium exporters today are Australia,
Kazakhstan, Russia, Canada, and Niger.) Higher prices could also justify
further
exploration to grow reserves.
In short, the TAP reactor enables known uranium reserves to be mankind's long-
term solution to an abundant, cheap supply of clean electricity.
Waste Stream
The TAP reactor greatly reduces waste as compared to conventional LWRs,
whether it is running on SNF or low-enriched fresh fuel. Figure 9 shows the
time
evolution of the actinides present in the TAP reactor starting from an initial
load of SNF.
As shown, the majority of the isotopes remain essentially in a steady state
across many
decades. The increases in U-236 and Pu-240 are welcome from an anti-
proliferation
17
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
standpoint, because these isotopes tend to capture neutrons in a nuclear
weapon, retarding
detonation.
A 520 MWe light-water reactor would contain approximately 40 tons of fuel and
generate 10 metric tons of SNF each year. The SNF contains materials with half-
lives on
the order of hundreds of thousands of years. Although reprocessing methods are
available
for partially reducing the waste mass, they are currently cost prohibitive and
accumulate
pure plutonium as a byproduct.
A basic mass flow and waste composition for a 520 MWe TAP reactor are as
follows: The reactor starts with 65 tons of actinides in its fuel salt. Each
year, 0.5 tons of
fission products are filtered from the system and a fresh 0.5 tons of fuel is
added, keeping
the fuel level steady. At reactor end of life, the inventory of fuel remaining
in the reactor
may be transported for use in another TAP reactor. Alternately, it may be
casked and
stored in a repository.
A breakdown of the methods and approximate quantities removed per year by one
520 MWe plant is shown in Table 1.
Gases: The fission products krypton and xenon are removed in the form of a
gas,
via an off-gas system, and are compressed and bottled on site. Trace amounts
of tritiated
water vapor are removed and bottled via the same process. A small fraction of
the noble
fission products are removed directly via the off-gas system.
Solids: Noble and semi-noble metal solid fission products, as well as other
species
that form colloids in the salt, are removed from the salt as they plate out
onto a nickel
mesh filter located in a sidestream in the primary loop.
Dissolved lanthanides: While they are less serious factors than krypton and
xenon,
it is desirable to remove lanthanides from the fuel salt for best operation.
We have several
options here. Our current approach is to remove lanthanide fission products
via a liquid-
metal/molten salt extraction process being developed by others in the USA and
France.
This process can ultimately convert the dissolved lanthanides into an oxide
waste form.
This waste form is fairly well understood, because spent nuclear fuel from
LWRs is in
oxide form. This oxide waste comes out of the processing facility in ceramic
granules and
can be sintered into blocks or any other form convenient for storage.
18
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Table 1. Fission product removal methods and approximate average removal rate.
Adapted in part
from [14].
Fission Product Removal Process Approximate Waste Form
removal rate, kg
per year
Kr, Xe, tritiated Helium sparging via 100 Compressed,
water vapor off-gas bottled gas
Zn, Ga, Ge, As, Se' Plating and filtration,
Nb, Mo, Ru, Rh,
some removal via off- 200 Metallic
Pd, Ag, Tc, Cd, In,
Sn, Sb, Te gas
Zr
Ni, Fe, Cr
Np, Pu, Am, Cm
(trace) Molten salt / liquid
Y, La, Ce, Pr, Nd metal extraction 200 Solid oxides
Pm, Gd, Tb, Dy,
Ho, Er, Sm, Eu
Sr, Ba, Rb, Cs
Compared to a similarly-sized light-water reactor, the annual waste stream is
reduced from 10 to 0.5 metric tons ¨ which is 95% less waste. Furthermore, the
vast
majority of our waste stream ¨ the lanthanides, krypton, xenon, tritiated
water vapor,
noble metals, and semi-noble metals ¨ has a relatively short half-life decay,
on the order
of a few hundred years or less. We believe mankind can tractably store waste
materials
on these timescales, compared to the hundreds of thousands of years required
for waste
from LWRs.
Of the 200 kilogram lanthanide mass removed by liquid metal extraction, we
estimate that approximately 20 kilograms will be actinide contaminant with a
longer half-
life similar to SNF. It may be most practical to leave such a small quantity
embedded in
the ceramic granules, as it would be well distributed and would not materially
extend the
time for the overall waste form to reach background levels. If desired,
however, the
actinides can be further separated with additional post-processing techniques.
In summary, compared to a light-water reactor, the TAP reactor emits 95% less
waste, with an overall waste storage time of a few centuries instead of
hundreds of
thousands of years.
19
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Better Inherent Safety
Molten salt reactors are a win for public safety. The main concern in a
nuclear
emergency is to prevent wide-spread release of radioactive materials. The TAP
reactor's
materials and design greatly reduce the risk of reactor criticality incidents,
shrink the
amount of radioisotopes in the primary loop, eliminate driving forces that can
widen a
release, and provide redundant containment barriers for defense in depth.
Self-Stabilizing Core
Like light-water reactors, molten salt reactors have a strong negative void
coefficient and negative temperature coefficient. In molten salt reactors,
these negative
coefficients greatly aid reactor control and act as a strong buffer against
temperature
excursions. As the core temperature increases, the salt expands. This
expansion spreads
the fuel volumetrically and slows the rate of fission. This stabilization
occurs even
without operator action and does not require control rods to function.
Control rods are included in our design to aid in power-up and can be used to
SCRAM the core. Molten salt reactors, however, are operator-controlled
primarily via the
turbine and not by control rods. Slowing the turbine extracts less heat from
the salt,
thereby increasing its temperature, which in turn decreases reactivity. Once
the reactor
reaches the lower power level where heat produced is equal to the turbine heat
draw, the
system re-stabilizes. It is not possible to have a runaway reaction due to
increasing the
cooling level too rapidly via the turbine ¨ drawing too much heat from the
core too
freezes the salt. These dynamics provide tight negative feedback loops and
give the
system inherent stability.
Although the TAP reactor is meant for baseload operation, the ability to
control
heat output via the turbine enables load following operation.
Smaller Inventory of Radionuclides
As shown in Table 2, a typical 1 GWe light-water reactor core has an inventory
of
2 to 7 tons of radionuclides that may conceivably escape during accident
conditions. By
convention, these core inventory numbers do not include uranium.
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
These are core inventories that are used to calculate source terms for
radionuclude
releases in various accident scenarios. However, some accidents such as
Fukushima
extend to the SNF pool. If a large SNF pool is assumed, then the total plant-
wide
radionuclide inventory may exceed 30 tons.
A 520 MWe TAP reactor maintains far less source material on hand, because it
is
much more fuel-efficient than an LWR. Furthermore, noble gases, noble metals,
and
lanthanides are removed continuously from the system, as shown previously in
Table 1.
Our radionuclide inventory is therefore just 0.9 tons in a 520 MWe reactor,
which is
significantly less than what would be present in a similarly-sized light-water
power plant.
This reduction shrinks the maximum size of a potential release.
Table 2. Radionuclide inventories (normalized to 100 MWe, net generation) in
the
primary loop for BWR, PWR, and TAP reactor accident analyses. BWR and PWR
numbers, chemical groups, and elements in the groups are adapted from [15].
Following
[15], LBU indicates an average burnup of 28 GWd per MTHM and HBU indicates an
average burnup of 59 GWd per MTHM.
Table 2. Radionuclide inventories (normalized to 100 MWe. net generation) in
the primary loop for
BWR, PWR, and TAP reactor accident analyses. BWR and PWR numbers, chemical
groups, and
elements in the groups are adapted from [15]. Following [15], LBU indicates an
average burnup of 28
GWd per MTHM and HBU indicates an average burnup of 59 GWd per MTHM.
Peach Bottom Unit 3 Sequoyah
Unit 1 (1148
Elements in the (1138 MWe BWR), MWe
PWR), TAP Reactor
Chemical Group (520 MWe M512),
Group kg per 100 MWe kg per 100
MWe
kg per 100 MWe*
LBU HBU LBU HBU
Noble Gases Kr, Xe 32 77 26 45
<0.1
Halogens Br, I 1 3 1 2
<0.1
Alkali Metals Rb, Cs 18 44 14 25 3
Tellurium Group Se, Sb, Te 3 7 2 4 <0.1
Alkaline Earths Sr, Ba 14 33 11 19 8
Noble Metals Co, Mo, Tc, Ru,
44 112 18 32 <0.1
Rh, Pd
Y, Nb, La, Pr, Nd,
Lanthanides**
Pm, Sm, Eu, Am, 43 109 34 61 22
Cm
Cerium Group Zr, Ce, Np, Pu 106 201 85 126 137
Total
261 586 191 314 170
(kg per 100 MWe)
Total
2968 6665 2196 3600 884
(kg in Entire Plant)
* Steady-state values in the primary loop, assuming fission product removal as
described above.
21
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
** By convention in NUREG-1246, Cm and Am are placed in the lanthanide group.
Reduced Driving Force
As described in some detail in our comparison of solid and liquid fuels, light-
water reactors can experience enormous driving forces during accident
scenarios. These
forces can come from a hydrogen explosion, a steam explosion, or in some
reactors, a
high system pressure of 150 atmospheres.
The chance of a high driving force is greatly reduced in a molten salt
reactor,
because it operates at near-atmospheric pressures, and there is little chance
of a vapor
explosion. The highest pressure element is the steam turbine. Nuclear reactors
already
protect against an upstream pressure transient ¨ such as a turbine break ¨
using rupture
disks, a passive safety feature that reduces system pressure without any
external action
required. We adopt the same approach to protect the nuclear island in the TAP
reactor.
Passive Safety and Inherent Resistance to Beyond-Design-Basis Events
A significant vulnerability common to all currently operating commercial light-
water reactors is that they require a continuous supply of electricity to pump
coolant over
their core to prevent a meltdown. By definition, a passively safe nuclear
reactor is one
that does not require operator action or electrical power to shut down safely
in an
emergency. It is a further goal that the reactor be able to safely cool during
a station
blackout without any outside emergency measures. An inherently safe reactor
will be
able to achieve these goals even in the face of an unanticipated or beyond-
design basis
event.
No reactor design assures perfect safety. However, the TAP reactor is a major
advance over light-water reactors because it is passively safe (primarily due
to its freeze
valve) and can passively cool its drained core via cooling stacks connected to
its auxiliary
tank, as described above. If the freeze valve fails, the control rods may be
inserted by
operator action or passively via an electromagnetic failsafe, thereby making
the reactor
subcritical. If the control rods or other active measures cannot be used, the
hot fuel salt
will simply remain in the reactor vessel. Heat will cause the salt to expand,
thereby
reducing reactivity. If the freeze valve fails and the salt continues to
increase in
22
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
temperature, the zirconium hydride moderator rods will decompose. The lack of
neutron
moderation brings the reactor to a sub-critical state.
If the salt increases in temperature enough to induce material failure in the
vessel,
then the salt will flow via gravity into a catch basin, shown in Figure 2,
located
immediately below the vessel. The catch basin in turn drains via gravity into
the auxiliary
tank. The reactor and its catch basin are sealed within a concrete chamber
only accessible
by hatch. Thus, even in this worst-case accident scenario, the system is
confined, non-
flammable, and shuts down passively.
If fuel salt through some further circumstance escapes the primary containment
surrounding the primary loop, it will still be inside the concrete secondary
containment
structure, which is located at least partially below grade. An intermediate
loop creates a
buffer zone between the radioactive materials in the reactor and the non-
radioactive water
in the steam turbine. The steam is at a higher pressure than the intermediate
loop and the
intermediate loop is at a higher pressure than the primary loop, so that any
leaks in heat
exchangers will cause a flow toward the core rather than out of the core. Any
small
counter-pressure flow across the primary heat exchanger is trapped in the
intermediate
loop. The intermediate loop feeds into a steam generator, and both are also
within the
concrete secondary containment structure. If the fuel salt, despite all
existing safety
mechanisms in the system, escapes the containment structure, it will return to
solid form
once it cools below approximately 500 C.
Table 3 summarizes how fundamental material choices affect key safety aspects
for light-water and TAP reactors. TAP reactors have greater inherent safety,
which is
particularly important for unanticipated and beyond design-basis accidents.
Table 3. Inherent Safety for Light-Water and TAP Reactors
1 GWe LWR 520 MWe TAP
Negative Void Yes Yes
Coefficient
Negative Temperature Yes Yes
Coefficient
Moderator Failsafe Water drains or boils off Moderator rods lose
function at
high heat due to marginal loss
of hydrogen
23
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Radionuclide 2-30 tons onsite <1 ton onsite
Inventory
Driving Force / 150 atmospheres 1 atmosphere
System Pressure
Driving Force / Peak fuel temperature is Peak fuel temperature is
500 C
Coolant 1900 C above coolant below boiling point; wide
safety
boiling point; steam margin
explosion risk
Driving Force / Peak fuel temp is 800 C Peak fuel temperature is
500 C
Runaway Exothermic above exothermic generation below exothermic generation
Hydrogen Generation point; fire explosion risk point; wide safety margin;
no
water in core
Table 4 compares the physical barriers for a light-water reactor and a TAP
reactor. The TAP reactor has no fuel cladding because it uses liquid fuel.
Auxiliary
support to the vessel and cooling boundary is provided by a passive freeze
plug, which
drains the fuel from the vessel into an underground auxiliary tank during
emergency
conditions. An additional boundary is provided around the vessel and cooling
system
with a catch basin and an intermediate cooling loop.
Table 4. Physical Berner Comparison
LWR TAP
Fuel Material Barrier Oxide matrix Salt carrier solidifies
<500 C
Cladding Barrier Zirconium cladding
Vessel and Cooling Stainless steel vessel and Hastelloy-N vessel
and heat
Boundary heat exchanger exchanger
Auxiliary Tank Freeze plug passively
drains
fuel to underground auxiliary
tank
Primary Containment Yes Yes
Structure
Catch Basin and Yes
Intermediate Loop
Secondary Containment Yes Yes
Structure
Exclusion Zone Yes Yes
In sum, today's nuclear plants are designed such that an explosion or steam
rupture could have wide area consequences, but safety is assured
probabilistically
24
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
through the use of multiple independent systems of redundant function, adding
cost and
complexity. TAP reactors draw on these redundant system techniques in places,
but we
ultimately provide a more resilient safety foundation ¨ molten salt is
inherently less
capable of a wide-area public disaster.
Reactor Cost
There are a range of commercial power plants that can be envisioned using
Transatomic Power's technology. We worked with Burns & Roe, an experienced
nuclear
engineering, procurement, and construction firm, on a system-wide pre-
conceptual plant
for a 550 MWe (gross generation) TAP reactor, with a net output of 520 MWe.
Such a plant would serve a gap in the market ¨ today's most modern light-water
reactors are typically large units aimed at 1000 MWe and above; a recent push
to develop
small modular reactors (SMRs) is aimed primarily at 300 MWe and below. The 520
MWe size may be particularly attractive to utilities because it is sized
similarly to aging
coal plants. The overnight cost for an nth-of-a-kind 520 MWe size was
estimated at $2.0
billion with a 3-year construction schedule.
The TAP reactor can realistically achieve these overnight costs because the
outlet
temperature of 650 C allows for higher thermal efficiency than current LWR
temperatures of 290-330 C, enabling a significant savings in the turbine and
balance of
plant. There are additional savings because (1) the reactor and heat transfer
equipment
operate near atmospheric pressures, reducing complexity and expense for both
equipment
and structures; and (2) the TAP reactor does not need onsite SNF underwater
storage with
its associated water treatment, leak detection, backup water, and backup
generator
systems.
There are several cost disadvantages for the TAP reactor that were anticipated
in
this analysis as well. We need to keep our piping warm to prevent salt freeze-
outs. We
must contend with tritiated water vapor capture at high temperatures. We use
an
intermediate loop filled with non-radioactive salt to separate the steam cycle
from the
fuel-salt. We also require structural space for fission product removal.
Nevertheless, the
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
analysis shows these cost additions are greatly outweighed by the savings
described
above.
The $2 billion price point can greatly expand the demand for nuclear energy,
because it is a lower entry cost than large-sized nuclear power plants, which
are usually
well above $6 billion and take longer to construct than the smaller TAP
reactor. A lower
price for a smaller unit will expand the number of utilities that can afford
to buy nuclear
reactors, better match slow changes in demand, allow greater site feasibility,
and reach
cashflow breakeven faster. The speed of construction and faster payback also
reduce
financing costs.
TAP reactors will also deliver a low levelized cost of electricity (LCOE).
While
most observers assume nuclear fuel costs are near zero, the Nuclear Energy
Institute
estimates the 2011 cost was actually 0.68 cents per kilowatt-hour. As the
above fuel cycle
figures illustrate, we expect to produce far more electricity per ton of ore
than the current
fuel cycle, driving these costs down toward zero. The TAP reactor is refueled
continuously for a high uptime. Finally, the 520 MWe size will absorb
overheads better
than smaller SMRs.
Lowering the Hurdles for a U.S. Repository
The United States has set aside a $30 billion trust for a repository and has
64,000
tons of SNF to store ¨ approximately $500 per kilogram of SNF. However, our
country
has not been able to agree on a location or final design for the repository.
Why not reprocess? The cost to reprocess as the French do is about $1,000 per
kilogram of SNF, which is well above what is available in the U.S. Waste
Disposal Trust
Fund. Meanwhile, SNF can be held inside existing wet storage pools at near-
negligible
cost. As pools fill up, SNF older than 3-10 years can be dry casked for
roughly $100 per
kilogram and stored for up to 40 years, making this method a cost-effective
stopgap.
About one-quarter of US SNF has been dry-casked. The other 48,000 tons remain
in wet
pools, adding to the plant inventory of radionuclides described in Section
3.2.
The TAP reactor can use fresh uranium fuel or SNF. Utilities can buy fresh
uranium from commercial suppliers. The business case for a utility using SNF
is
26
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
somewhat more complicated, because the SNF requires additional handling costs
as
compared to fresh fuel. The plant must (1) transport and receive the
radioactive spent fuel
rods, (2) remove the cladding physically, and (3) dissolve the uranium oxide
into the
molten salt or convert it to a gas that can be injected into the molten salt.
The techniques
are well known because the same three initial steps must be employed in
reprocessing
plants such as at Le Havre in France or similar facilities existing at the
Idaho National
Laboratory [8]. We avoid, however, all of the remaining chemical steps that
are the main
cost drivers of the work. If reprocessing costs $1000 per kilogram, we could
potentially
perform just the initial steps for a fractional amount, perhaps in a small
number of
regional facilities that ship fuel directly to TAP reactors. Our initial
assessment is that a
disposal charge of $500 per kilogram of SNF is achievable, affordable, and
more cost-
effective than reprocessing and would be within the budget allowed by the U.S.
Waste
Disposal Trust Fund.
The existing 64,000 tons of SNF contain an enormous amount of energy. If all
U. S . light-water plants were replaced tomorrow by TAP reactors, it would
still take 350
years to consume all of the existing SNF. Even if we expand the role of
nuclear by also
converting all coal plants to TAP reactors, we could still run for 150 years.
The SNF
needs to be stored in the meantime. Furthermore, the TAP reactors would
themselves
create small amounts of waste to store. We therefore cannot use TAP reactors
to avoid a
U.S. repository entirely. TAP reactors do, however, allow us to build a
smaller and
simpler repository. SNF would only need to be stored for a few hundred years
instead of
hundreds of thousands of years. Furthermore, by avoiding a great deal of
future SNF, we
may avoid the need to build a second or third repository.
Anti-Proliferation Analysis
The TAP reactor represents a major victory for non-proliferation, because it
cuts
future production of SNF while slowly reducing SNF stockpiles from the past.
Today, the world's main tool to block plutonium proliferation is to guard
irradiated materials. Light-water reactors are, however, a troubling
contributor to the
problem. One ton of SNF contains enough Pu-239 for one atomic bomb [16], and
the
27
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
world has accumulated 270,000 tons of commercial SNF. This figure is growing
by some
10,000 tons per year, and is further accelerating as the rest of the world
builds more light-
water nuclear power plants in more countries. Starting up a typical 1 GWe
light-water
reactor in a foreign country requires 90 tons of initial fuel, and a further
20 more tons of
fuel, on average, for each year that the reactor is in operation. After 60
years, the foreign
country has 1200 tons of SNF ¨ enough for a weapons program to build over one
thousand atomic bombs. The foreign SNF must therefore be guarded in
perpetuity, and it
is forever a threat to become the materials source for a weapons arsenal if
the state goes
rogue or if the material is stolen.
Our design is proliferation resistant because no process preferentially
removes or
extracts any isotope, and the facility does not enrich source material. We do
not separate
pure uranium or pure plutonium or any precursor of pure uranium or plutonium.
The
source material is at high temperature and diluted across the molten fluoride
salt, making
theft impractical.
There are three separate waste streams emerging from the TAP reactor. The
first
is from a continuously-operating off-gas system that removes contaminants,
including
fission products, fission product daughters, water, oxygen, and small amounts
of tritiated
water vapor, from the primary loop. The second waste stream is composed of the
noble
and semi-noble metals that plate onto a mesh filter located in the primary
loop. Neither
contains any source material useful for atomic weapons.
The third waste stream is made up of lanthanide fission products. We remove
these fission products using molten salt/liquid-metal extraction, a process
under
development by others in France and the USA. We use this method because it is
highly
effective at removing lanthanides with minimal actinide contaminants in the
waste
stream, and never separates pure plutonium or uranium. Furthermore, most of
the
separation steps occur in counter-flow columns that would be complex to
modify. The
two final steps use electrochemistry: one removes minor actinides from a
liquid metal
stream, and the other removes lanthanides from the liquid metal stream. As
discussed
previously, the lanthanide waste stream ultimately emerges an as oxide that
can be
sintered into blocks or other solid shapes suitable for storage.
28
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
Despite the efficiency of the process, the lanthanide waste stream of 200
kilograms per year is contaminated by detectable levels of actinides,
approximately 20
kilograms total, including small amounts of uranium and plutonium. The uranium
contaminant is at 1.8% enrichment, and is therefore not a proliferation
concern. Less than
0.1% of the lanthanide waste stream is plutonium contaminants ¨ a factor of 10
reduction
compared to LWR spent fuel, which is approximately 1% plutonium. The
lanthanide
fission product waste stream would therefore not be a practical source of
weapons
materials for a rogue state.
Finally, we note that the several countries are currently struggling to handle
their
stockpiles of plutonium. Plutonium is isolated as a by-product during the
reprocessing
techniques used in France, the UK and elsewhere. Due to the versatility of
molten salt
reactors, future TAP reactors could burn this plutonium after it is
downblended and
mixed with natural uranium. Directly reducing stockpiles of weapons plutonium
is a
significant anti-proliferation benefit.
Comparison to Other Waste-Burning Reactors
Several advanced fast reactor concepts have also been proposed to burn waste.
However, fast reactors have proven difficult to scale up despite major past
investments.
All fast reactors are challenged by high neutron fluence ¨ an order of
magnitude higher
than traditional reactors ¨ and the resulting damage that occurs to vessels
and equipment.
Fast reactors also face proliferation concerns because they can produce excess
plutonium during operation. Some fast reactors handle this issue by sealing
the reactor so
that there is no external access to the core, but this lack of access
increases the materials
challenges of the design even further. Additionally, some fast reactors have a
fire risk due
to their sodium metal coolant. Molten salt does not have this risk. Molten
salt reactors
can also be built at considerably lower cost than gas fast reactors.
The TAP reactor aims to close the fuel cycle with a commercially viable and
scalable technology. We use a thermal spectrum, which reduces component damage
as
compared to a fast reactor, and we achieve greater inherent safety for the
public. The
fundamental principles of the design have already been demonstrated at the Oak
Ridge
29
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
National Laboratory. We modify this previous design to yield exciting benefits
without
demanding dramatically new materials. Our improvements can also be
demonstrated at a
small scale, reducing development costs. For these reasons, the TAP reactor is
the best
and most practical concept for closing the nuclear fuel cycle.
Why Not Thorium First?
The TAP reactor's primary innovations ¨ a novel combination of moderator and
fuel salt ¨ can also be adapted for use with thorium. Transatomic Power
believes that the
thorium fuel cycle holds theoretical advantages over uranium in the long run
due to its
generally shorter half-life waste, its elimination of plutonium from the fuel
cycle, and its
greater natural supply. However, we chose to start with uranium for several
reasons: (1)
there is a great deal of spent nuclear fuel, and we want to harness its energy
while
reducing the risk of onsite SNF storage; (2) the industry already has a
commercial fuel
cycle developed around uranium; (3) we already greatly eliminate waste; and
(4) we
already greatly expand the energy potential of existing uranium supplies.
Thorium reactors do not contain plutonium, but they do have a potential
proliferation vulnerability due to the protactinium in their fuel salt.
Protactinium has a
high neutron capture cross section and therefore, in most liquid thorium
reactor designs,
it must be removed continuously from the reactor. The process for doing this
yields
relatively pure protactinium, which then decays into pure U-233. By design,
the pure U-
233 is sent back into the reactor where it is burned as its primary fuel. The
drawback,
however, is that U-233 is a weapons-grade isotope that is much easier to
trigger than
plutonium. It is possible to denature the U-233 by mixing it with other
uranium isotopes,
or modify the design to further reduce diversion risk, but further research is
required to
implement these anti-proliferation measures in thorium molten salt reactors.
Future Advances
The basic TAP reactor design described in this report will benefit from future
innovations in a number of different ways. Improvements to complementary
technology
will become commercially available over time. These technologies include high
temperature ceramics such as SiC-SiC composites for heat exchangers and other
reactor
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
internals, which will allow us to increase the reactor's operating temperature
and increase
thermal efficiency. We will likely be able to incorporate closed loop Brayton
cycles once
that technology becomes readily commercially available.
As renewables grow more prevalent and grid supply becomes more variable, we
may also adapt the plant for better load-following. Molten salt reactors are
inherently
better able to load-follow than solid-fueled reactors, because the off-gas
system prevents
the neutron poison xenon from building up in the primary loop. In solid-fueled
reactors,
decreasing the power level causes an increase in xenon, because xenon is not a
direct
fission product. Following shutdown, light water reactors require on the order
of several
days for the xenon to decay enough to allow for restart. Boiling water
reactors and
advanced boiling water reactors are capable of overnight load following, but
this xenon
instability can reduce their load following performance by inducing local
power peaking
in the core. Molten salt reactors do not experience xenon instability, because
the off-gas
system quickly removes xenon from the primary loop, regardless of power level.
Other small modular reactor designs are capable of a crude type of load
following
via the following scheme: the power plant consists of an array of reactors in
the range of
50 ¨ 200 MWe, and the individual units are turned off and on depending on
power
demand. A major drawback of this system is that the multiple stop and restart
cycles may
damage the reactor components. In contrast, molten salt reactors like the TAP
reactor are
capable of much more precise and continuous load following.
These technology advances present bright new opportunities for nuclear power.
Reliable load following will allow reactors to adapt to daily and seasonal
changes in
electric demand and take advantage of the corresponding fluctuations in
electricity prices.
Furthermore, increasing the operating temperature of the plant will allow
these reactors to
expand into new markets such as process heat and synthetic fuel production.
Conclusions
Transatomic Power's molten salt reactor generates clean, passively safe, and
low
cost nuclear power from SNF or low-enriched fresh uranium fuel. The most
significant
differences between this reactor and previous molten salt designs are our
zirconium
31
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
hydride moderator and LiF-(Heavy metal)F4 fuel salt, which allow us to achieve
a very
high actinide burnup in a compact, cost-effective design.
Previous experimental work in conjunction with the TRIGA and SNAP reactors
has shown that zirconium hydride is stable at the temperatures and neutron
fluxes present
in Transatomic Power's reactor. Other experimental work at the Oak Ridge
National
Laboratory demonstrated the compatibility of modified Hastelloy-N with molten
fluoride
fuel salts.
The reactor has a thermal spectrum, which reduces neutron damage to the
moderator and other plant components as compared to a fast spectrum, and
consequentially lowers the costs associated with component replacement. There
are,
however, sufficient epithermal and fast neutrons to break down actinides. The
reactor is
highly proliferation resistant: it requires minimal fuel processing, and never
purifies
special nuclear materials. Furthermore, this plant possesses the appealing
safety benefits
common to most molten salt fueled reactor designs. It does not require any
external
electric power to shut down safely.
The TAP reactor solves some of the most pressing problems facing the nuclear
industry ¨ safety, waste, materials proliferation, and cost ¨ and can allow
for more
widespread growth of safe nuclear power.
Other embodiments
A number of embodiments have been described. Nevertheless, it will be
understood that various modifications may be made without departing from the
spirit and
scope of the disclosure. For example, these concepts can be applied to a
molten salt
reactor whose core is comprised of multiple zones with varying moderator and
fuel-salt
volume fractions. The purpose of the multi-region core is to increase the
conversion ratio
(as compared to a core with a uniform moderator volume fraction) while
maintaining
criticality.
In some implementations, the moderator is comprised of zirconium hydride and a
cladding to separate the moderator from the fuel-salt. Zirconium hydride is a
very
efficient moderator, meaning that it can create a thermalized neutron energy
spectrum
32
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
with a smaller volume than most other moderators. Lithium fluoride actinide
fluoride has
the advantage of having a higher actinide solubility than most other fuel
salts. This
combination of moderator and fuel salt enables criticality with a smaller core
volume
than typical molten salt reactors.
In other implementations the moderator may be graphite, beryllium oxide, metal
hydrides, or metal deuterides like zirconium deuteride, amongst others, or any
combination of two or more of these moderators. The solid moderator may be in
the form
of rods, annular rods, finned rods, wire-wrapped rods, spheres or pebbles,
large blocks
with fuel-salt channels going through the block, plates, assemblies of plates,
or any other
suitable geometry, or any combination of suitable geometries.
In some implementations, the fuel-salt is comprised of lithium fluoride and
actinide fluorides, where actinide fluorides can be a combination of actinide
elements, as
long as the fuel-salt includes at least one fissile isotope. In other
implementations, the
fuel-salt may be comprised of actinide fluorides, lithium fluorides, beryllium
fluorides,
zirconium fluorides, amongst others, or any combination of two of more these
salts.
Moderated regions are typically designed to maximize reactivity, which is
defined
as the positive or negative deviation of the multiplication factor (k) from
criticality, which
occurs when k = 1. Figure 10 illustrates how the multiplication factor varies
as a function
of moderator-to-fuel-salt volume fraction in one implementation using a
lithium fluoride
and actinide fluoride fuel-salt and a zirconium hydride moderator. This figure
was
generated from simulation of an infinite lattice of fuel-salt and moderator.
Pitch is the
center-to-center spacing between adjacent rods of moderator. The simulations
were
performed with MCNP6.
The conversion ratio is typically defined as the ratio of the rate of fissile
production to the rate of fissile loss. When the conversion ratio equals one,
the rates of
fissile production and destruction are exactly equal. In a simplified molten
salt reactor
system with a conversion ratio equal to one, the fissile concentration can be
kept constant
over time by continuously feeding a stream of fertile nuclei into the reactor
at a rate equal
to the rate of fission. (This and subsequent examples assume that all fission
products are
immediately removed from the system.) If the conversion ratio is greater than
one, the
33
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
fissile concentration will increase over time if fertile nuclei are
continuously fed into the
reactor. When greater than one, the conversion ratio is called the breeding
ratio. If the
conversion ratio is less than one, the concentration of fissile nuclei will
decrease over
time if only fertile nuclei are fed into the reactor. However, if enriched
uranium, for
example, is fed continuously into the simplified reactor system, the fissile
concentration
in the reactor will remain approximately constant if the fissile content of
the feed (ffeed)
is equal to one minus the conversion ratio (CR):
ffeed = 1 ¨ CR
The burnup (B), or fraction of the actinide fuel that is fissioned, can be
approximated
with the equation:
E
B= ____________________________________________
(1¨ CR)
where the E is the effective enrichment, or percentage by weight of fissile
nuclei
in the actinide fuel. Figure 11 shows that to achieve a high burnup, the core
must have a
high conversion ratio or high enrichment.
The conversion ratio varies as a function of fuel-salt and moderator volume
fractions. Figure 12 illustrates how the conversion ratio varies as a function
of fuel-salt
volume fraction in one exemplary implementation. In this example, the entire
volume is
comprised of either fuel-salt or moderator, so the moderator volume fraction
is equal to
one minus the fuel-salt volume fraction.
By looking at Figure 10 and Figure 12, one can see that the conversion ratio
is
highest where the entire core volume is fuel-salt and no solid moderator is
present.
However, the multiplication factor is greatest when the ratio of fuel-salt to
moderator is
approximately one, meaning there are approximately equal volumes of fuel-salt
and solid
moderator present in the core. The disclosed reactor incorporates within the
core multiple
distinct regions with varying volume fractions of solid moderator such that
the
conversion ratio of the combined regions is greater than that of a core
comprised of a
uniform lattice of solid moderator and fuel-salt while maintaining a
multiplication factor
equal to or greater than one.
34
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
One exemplary embodiment, illustrated in Figure 13, is comprised of a central,
moderated region surrounded by an outer, unmoderated region. The inner region
has fuel-
salt and solid moderator volume fractions at or near the combination that
maximizes the
multiplication factor. Figure 10 shows that reactivity is maximized when fuel-
salt and
solid moderator volumes are approximately equal. Therefore, the central,
moderated
region of this embodiment is comprised of equal volumes of fuel-salt (lithium
fluoride,
actinide fluoride) and solid zirconium hydride moderator. The outer region is
unmoderated (in that it does not substantially contain any solid moderator).
The addition
of the outer, unmoderated region decreases the multiplication factor of the
core, but also
increases the overall conversion ratio of the combination of the two regions.
Preliminary analyses with MCNP6 and SCALE 6.1 indicate that a core as
depicted in Figure 13, with a 2 meter diameter central moderated zone (50%
moderator,
50% fuel-salt) and a 0.5 meter thick unmoderated region, can achieve a
conversion ratio
of approximately 0.9 while maintaining a multiplication factor greater than
one.
Improvements to the conversion ratio are likely possible by increasing the
total diameter
of the core while also increasing the volume of the unmoderated zone relative
to the
moderated zone.
Other embodiments may be comprised of a central unmoderated region and an
outer, moderated region. Additional embodiments may be comprised of two or
more
regions, with at least two distinct volume fractions of fuel-salt and solid
moderator.
Figure 14 illustrates a variation of a two region core, with the unmoderated
region
in the center and surrounded by the moderated region. This configuration may
offer a
higher conversion ratio than the core in Figure 13, because the higher scalar
neutron flux
in the center of the core may increase the rate of fertile-to-fissile
transmutation.
Figure 15 expands upon this concept by adding a second unmoderated region
along the periphery of the core. The outer unmoderated region acts as a
neutron-
absorbing blanket that increases overall conversion ratio, reduces neutron
leakage out of
the core, and reduces neutron fluence and damage to the vessel wall. Increased
neutron
absorption in the outer unmoderated region is caused primarily by the
increased
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
concentration of U-238, which is a strong neutron absorber in the epithermal
energy
range.
The incorporation of a central unmoderated region, while increasing the
overall
conversion ratio of the core, also causes a decrease in the multiplication
factor. To reduce
the detrimental effect on the multiplication factor, the central region can be
designed to
have volume fractions of fuel-salt and moderator between fully unmoderated to
the
configuration that maximizes the multiplication factor (approximately 50% fuel-
salt, 50%
moderator). Figure 16 illustrates one implementation of this design, which has
an outer
unmoderated region, and central slightly moderated region, and a moderated
middle
region.
Accordingly, other embodiments are within the scope of the following claims.
References
[1] P.N. Haubenreich, Molten Salt Reactor Experiment, ORNL-4449 (1970).
[2] US DOE Nuclear Energy Research Advisory Committee and the Generation
IV International Forum, A technology roadmap for generation IV nuclear energy
systems. GIF-002-00 (2002).
[3] L. Dewan, Molecular Dynamics Simulation and Topological Analysis of the
Network Structure of Actinide-Bearing Materials, MIT Doctoral Dissertation,
Department of Nuclear Science and Engineering (2013).
[4] 0. Benes, R.J.M. Konings, Molten Salt Reactor Fuel and Coolant,
Comprehensive Nuclear Materials. 3 (2012) 359-389.
[5] J. Buongiorno, J.W. Sterbentz, P.E. MacDonald, Study of solid moderators
for
the thermal-spectrum supercritical water-cooled reactor, Nucl Technol. 153
(2006) 282-
303.
[6] M. Simnad, The UZrHx alloy: Its properties and use in TRIGA fuel, Nucl.
Eng. Des. 64 (1981) 403-422.
[7] K. Moore, Phase relationships in the a + 6 region of the Zr-H system, J.
Nucl.
Mater. 32 (1969) 46-56.
36
CA 02925576 2016-03-24
WO 2015/094450
PCT/US2014/057655
[8] J. Huangs, B. Tsuchiya, K. Konashi, M. Yamawaki, Estimation of hydrogen
redistribution in zirconium hydride under temperature gradient, J Nucl Sci
Technol. 37
(2000) 887-892.
[9] N. Ponomarev-Stepnoi, V. Bubelev, E. Glushkov, V. Garin, N. Kukharkin, G.
Kompaniets, V. Nosov, E. Chunyaev, Estimation of the hydrogen emission from a
hydride moderator by measuring the reactivity and using mathematical
statistics, Atomic
Energy. 102 (2007) 87-93.
[10] K. Terrani, M. Balooch, D. Wongsawaeng, S. Jaiyen, D. Olander, The
kinetics of hydrogen desorption from and adsorption on zirconium hydride, J.
Nucl.
Mater. 397 (2010) 61-68.
[11] M.S. El-Genk, Deployment history and design considerations for space
reactor power systems, Acta Astronaut. 64 (2009) 833-849.
[12] R.C. Robertson, Conceptual Design Study of a Single-Fluid Molten Salt
Breeder Reactor, ORNL-4541 (1971).
[13] H. MacPherson, Molten Salt Reactors, Reactor Handbook. 4 (1964).
[14] D. Holcomb, G. Flanagan, B. Patton, J. Gehin, R. Howard, T. Harrison,
Fast
Spectrum Molten Salt Reactor Options (2011).
[15] D. Powers, M. Leonard, R. Gauntt, R. Lee, M. Salay, Accident Source Terms
for Light-Water Nuclear Power Plants Using High-Burnup or MOX Fuel, SAND2011-
0128.Albuquerque: Sandia National Laboratories (2011).
[16] T.B. Taylor, Nuclear safeguards, Annu. Rev. Nucl. Sci. 25 (1975) 407-421.
37
