All nuclear reactors built up to now have required refuelling with fissile fuel periodically.
And looking to Generation 1 fusion reactors, they will need topping up with both deuterium and tritium. They may have a blanket where tritium is produced but not enough to sustain operation without injecting additional tritium from a 3rd party.
Figure 1 shows various reactor technologies grouped by refuelling concept: fissile fuelled with repeated refuelling vs fertile fuelled with just an initial load of ‘kickstarter’ fuel. These are further subdivided according to fuel type.
The number of sides of the geometric figures (circle, triangle, etc) denote level of difficulty. Only reactors at the circle and triangle level have ever been built. All fission reactors requiring refuelling with enriched fuel or Pu are in the triangle category.
There have been very significant investments in Generation 1 fusion and solid fuel fast reactors hoping to leap across the chasm to the realm of reactors that only require refuelling with fertile fuel (right hand side of Figure 1).
The fuel cost of fissile fuel is typically several orders of magnitude larger than fertile fuel.
The inherent technical challenges to be solved to achieve fertile fuelled reactors are:
- achievement of excellent neutron economy with minimum leakage of neutrons;
- minimisation of counterproductive neutron absorption inside the core; and
- reduction of losses in reprocessing or when moving fissile material from blanket to centre.
In this article, we present a new thorium reactor design that is refuelled with fertile fuel (Th232) only. Figure 2 shows the concept, a triple fluid molten salt reactor core operating in the thermal spectrum, having a molten thorium salt as blanket, heavy water as moderator and a molten uranium salt as fuel salt. Ideally, fission only happens in the fuel salt, neutrons are absorbed by the blanket where Th232 is converted into U233 predominantly. Trace amounts of U232 and U234 are also generated in the blanket. A proprietary system is used to transfer all species of uranium from the blanket to the fuel salt periodically several times per operational hour. Each transfer is a few grams of uranium.
The reactor is designed to output 100 MW of heat and the outer diameter of the Onion Core® is 2.4 m.
There are two barriers between the hot salt and the unpressurised heavy water and graphite insulation foam is placed in the cavity. The transfer of heat from the hot salts to the water has been measured to be ~35 kW. The majority of the temperature increase in the heavy water originates from neutrons and gammas. 5 – 7% of the total reactor energy is injected into the heavy water and can be used for water desalination or district heating.
In order to be able to breed more new fissile fuel than is being consumed, we also need to remove the majority of fission products outside the lanthanides group.
Figure 3 shows how the Onion Core® is connected to pipes, pumps, heat exchanger and tanks.
The use of a heat exchanger creates three barriers between the radioactive salt and the salt which is sent outside the cocoon to the customer, eg, for electricity generation. The Cocoon wall and the Insulation wall are several meters thick to provide shielding from radiation.
When power to the pumps is cut they stop and all liquids (salt & water) drain into their respective dump tanks in 10-300 seconds depending on the volume of each circuit. Cutting the power to the pumps is the primary safety feature. Not all pumps are shown in Figure 3. There is one pump for each of the four channels in the Onion Core®.
This molten salt reactor is configured such that it can only load follow. This means that the reactor core can only produce the same amount of energy as is removed from the salt circuit leaving the left side of Figure 3. Control rods are not needed and the volume of heavy water in the inner water region can be adjusted to criticality.
This type of reactor cannot have a loss of coolant accident with subsequent core meltdown and it cannot have a rapid steam explosion of the type that can be envisaged for PWRs. It also does not have a spent nuclear fuel pool.
Figure 4 shows a 25 year burnup simulation of the reactor starting on a 5% enriched uranium fuel salt. It can be seen that the reactor consumes the majority U235 within the first 5 years, while it builds up U233 and plutonium in the fuel salt. It can be seen that ~70% of the power originates from uranium and ~30% originates from plutonium after 5 years. Gradually shifting towards ~85/15 % at 25 years. Every 5 years we siphon off fissile inventory (200 – 500 kg), which can be used to start other reactors. The simulation assumes lithium 7 (Li7) enriched to 99.999% purity and 99.9% deuterium content in the water. Neutron leakage from the reactor core is ~2%.
Planned tests
Thermal expansion, thermal cycling and thermal heat transfer have been tested in a full scale prototype non-nuclear rig in Copenhagen. The first nuclear test of the reactor is planned at Paul Scherrer Institute (PSI), Switzerland, in 2027. This first demonstration of criticality seeks to avoid delays by minimising complexity. Therefore, it does not have transfer of uranium from blanket to fuel salt. It does not have the SiC/SiC core envisaged for the commercial offering, and only some of the fission products are removed. For subsequent tests additional features will be added. However, with the first test at PSI, we will be able to validate the accuracy of the simulations to a point where uncertainties related to the postponed features become negligible.
Already the simulations carry substantial certainty that the Copenhagen Atomics reactor can cross the chasm to fertile fuelled reactors. But the test in 2027 will be the single event where most uncertainty is removed, between now and the point in time at which the first reactor attains 25 y of operation.
The uncertainties around the Copenhagen Atomics reactor concept mainly revolve around the use of SiC/SiC materials for core components, Li7 enrichment, online fission product removal and transfer of uranium from blanket to fuel salt. All these concepts are at a low TRL at time of writing, but are all expected to mature to commercial reactor deployment by 2030.
The limitation on the lifetime of the reactor is a combination of corrosion, salt compatibility, temperatures, stress/strain and radiation. It is currently the belief that the reactor core will be the limiting factor and that a life of 5 years can be achieved after a few iterations. A large percentage of the remaining work revolves around testing and approval of the materials used in critical components, such as reactor core, pump, pipes, heat exchangers and dump tanks.
A Copenhagen Atomics power plant is planned to last for 50+ years and many components are replaced every 5 years including the reactor vessel. As the technology matures these components may last longer than 5 years. An LWR replaces 300+ tonnes of materials per 5 years, Copenhagen Atomics power plants will replace less than that.
A new era
The Copenhagen Atomics reactor concept ushers in a new era of fertile fuelled reactors, promising orders of magnitude lower fuel cost and smaller fuel volume, smaller reactor core size and, overall, radically less expensive reactors.
These reactors create less waste, can easily be mass manufactured on assembly lines and transported on regular lorries and can even use transuranics from spent LWR fuel as kick starter fuel. It is likely that electricity prices of around
$20/MWh can be reached from 2035.
