ThEC13 Daniel Mathers daniel.p.mathers@nnl.co.uk The Thorium Fuel Cycle ThEC13 Daniel Mathers daniel.p.mathers@nnl.co.uk

Outline Content: Background Sustainability, proliferation resistance, economics, radiotoxicity Advantages and disadvantages Fuel Cycles

Fertile thorium A Thorium Fuel Cycle needs Uranium or Plutonium Th-232 is the only naturally occurring thorium nuclide It is a fertile nuclide that generates fissile U-233 on capturing a neutron Th-232 is fissionable in that it fissions on interacting with fast neutrons > 1 MeV kinetic energy Fertile conversion occurs with thermal neutron captures: Th-232 (n,g) Th-233 (b-) Pa-233 (b-) U-233 U-233 has a high thermal fission cross-section and a low thermal neutron capture cross-section The fission/capture ratio for U-233 is higher than the other major fissile nuclides U-235, Pu-239 and Pu-241 This is very favourable for the neutron multiplication factor and minimises the probability of neutron captures leading to transuranics A Thorium Fuel Cycle needs Uranium or Plutonium to initiate a fission reaction

Thorium fuel cycles Options for a thermal reactor are: Once-through fuel cycle with Th-232 as alternative fertile material to U-238 with U-235 or Pu-239 driver U-233 fissioned in-situ without reprocessing/recycle Modest reduction in uranium demand and sustainability Recycle strategy with reprocessing/recycle of U-233 Much improved sustainability analogous to U/Pu breeding cycle But some technical difficulties to overcome Options for a fast reactor are: MSFR and other Gen IV concepts (Sodium cooled fast reactors, ADS systems) All require U / Pu to initiate fission reaction Neutron balance in a thermal reactor For U-235 fissions average number of fission neutrons n ~ 2.4 When a neutron causes a fission it is removed from the system, which accounts for 1 of the 2.4 available neutrons Unavoidable neutron losses occur due to neutron captures in the fuel, structural materials, moderator, control materials and surface leakage and amount to about 0.7 to 0.8 neutrons in current thermal reactor designs Leaving about 0.6 to 0.7 neutrons available for fertile captures of U-238 to Pu-239 U-235 thermal reactors are therefore limited to a conversion ratio of ~0.6-0.7 and cannot operate as breeder reactors Thermal breeder The neutron balance for U-233 is more favourable (because neutron captures in U-233 are very low) and this gives the possibility of a conversion ratio close to 1.0 in a thermal reactor This is a breeding system Not quite ideal, but much better than other thermal reactors Not possible with U/Pu fuel cycle The possibility of a thermal breeder reactor was the driver for the thorium fuel cycle in the 1950s With uranium ore expected to be scarce Fast breeder technology more difficult to realise

Sustainability / Inherent Proliferation Resistance Thorium abundance higher than uranium Thorium demand lower because no isotopic enrichment Thorium economically extractable reserves not so well defined Rate of expansion of thorium fuel cycle will be limited by the slow conversion rate Inherent proliferation resistance U-233 is a viable weapons usable material High U-232 inventory implies high doses unless shielded Low inherent neutron source suggests that U-233 weapon design may be simplified and potentially more accessible U-233 fissile quality hardly changes with irradiation

Economics and Radiotoxicity U-233 recycle has lower demand on thorium than uranium because there is no isotopic enrichment process U-233 recycle potentially reduces the ore procurement cost and eliminates the enrichment cost Future uranium and thorium market prices unknown Short term economic barrier presented by need for R&D to demonstrate satisfactory fuel performance Radiotoxicity Spent fuel activity/radiotoxicity dominated by fission products for 500 years after discharge U/Pu long term fuel activity determined by activity of Np, Pu, Am and Cm Th/U-233 long term fuel activity has only trace quantities of transuranics and therefore lower radiotoxicity after 500 years However, this only applies to the long term equilibrium condition with self-sustained U-233 recycle In a practical scenario, the reduction in radiotoxicity is more modest than the long term equilibrium would indicate It is too soon to say whether the thorium fuel cycle will be economically advantageous Need to compare radiotoxicity over range of timeframes

Advantages of Th fuel cycle Thorium more abundant than uranium and combined with a breeding cycle is potentially a major energy resource Low inventories of transuranics and low radiotoxicity after 500 years’ cooling Almost zero inventory of weapons usable plutonium Theoretical low cost compared with uranium fuel cycle ThO2 properties generally favourable compared to UO2 (thermal conductivity; single oxidation state) ThO2 is potentially a more stable matrix for geological disposal than UO2

Void coefficient mitigation Supplementing a U/Pu recycle strategy Thorium fuels drive the void coefficient more negative in thermal and fast systems A positive void coefficient is an undesirable in-core positive feedback effect unless counteracted by other feedback effects In LWRs a positive void coefficient is usually considered unacceptable and limits the total plutonium load in MOX fuel to <12 w/o This is a potential restriction with poor fissile quality plutonium Thorium-plutonium fuel could allow significantly higher total plutonium loads (up to ~18 w/o), giving more flexibility for plutonium re-use in LWRs I have a yet to be published paper from Cambridge University which indicates that the void coefficient becomes more negative when U-Pu MOX is replaced with Th-Pu MOX. The paper examines Reduced Moderation PWR (RM-PWR), Reduced Moderation BWR (RM-BWR) and SFR and shows a more negative void coefficient in all of them. Mike Thomas has shown the same for PWR, a result which is consistent with AREVA’s work. THIS BIT IS JUST FOR MY OWN SATISFACTION: Why does thorium have this effect? Th-232 is a resonant nuclide with a large double resonance centred at about 22 eV (see slide at end of this pack). Voiding the water density in an LWR shifts the neutron spectrum towards this resonance and increases the proportion of slowing down neutrons that are removed by resonant captures. The same effect occurs with U-238 (very large and narrow lowest resonance at 6.5 eV), but the resonance peak in Th-232 is broader and lower than the lowest U-238 resonance, possibly leading to reduced self-shielding in Th-232 and thereby affecting the void coefficient. In other words, because U-238 is more heavily self-shielded, when voidage is introduced all the neutrons in the range of the lowest U-238 resonance are already removed by the resonance, the change in 1/E scattering source has limited impact. In Th-232 there is less self-shielding and therefore the change in 1/E scattering source is felt more strongly. EXTRACT FROM KIM & DOWNAR PAPER: One of the most important differences in the two isotopes is that the resonance integral of Th232 is about 3 times smaller than that of U238. This has an important implication for the void reactivity in a tight pitch lattice as will be discussed later. It is important to note that despite the large differences in the resonance integral, the effective resonance capture of both fertile fuels is comparable because resonance self -shielding reduces the effective resonance capture of U238. Thorium-based fuel designs also have the potential of enhanced safety characteristics in intermediate spectrum designs. Because of the reduced resonance integral and the higher fast fission threshold in Th232 compared to U238, the void coefficient of the thorium pins is larger and negative in the tight lattice cases A Possible way to manage plutonium stocks with poorer fissile quality and to allow time for thorium plutonium MOX qualification

Radiotoxicity Thorium-plutonium MOX fuel theoretically could be advantageous for UK plutonium disposition Detailed assessment by NNL of decay heat load and radiotoxicity per GWye shows there is only a marginal difference between Th-Pu MOX and U-Pu MOX This is a holistic calculation that accounts for the total decay heat outputs of different scenarios In the Th-Pu and U-Pu MOX cases, the decay heat is concentrated in the MOX assemblies, whereas in the UO2 reference case it is distributed over a larger number of UO2 assemblies

Decay heat Model uses standard PWR 17x17 assembly The Decay heat is per Gwye, therefore it is a normalised figure i.e. decay heat from assemblies (of these three types) irradiated in a reactor that produces 1GW of energy per year

Many technological issues to address - MSR is a long term option Molten salt reactor Molten Salt Reactor (MSR) Generation IV International Project is researching MSR Gen IV MSR will be a fast spectrum system Molten salt fuel circulates through core and heat exchangers On-line reprocessing to remove fission products Ideally suited to thorium fuel as fuel fabrication is avoided Equilibrium fuel cycle will have low radiotoxicity (fission products only) Many technological issues to address - MSR is a long term option

Accelerator driven system Accelerator driven system (ADS) Sub-critical reactor core Proton beam provides neutron source in spallation target Neutron source multiplied by sub-critical core

Disadvantages of Th fuel cycle Th-232 needs to be converted to U-233 using neutrons from another source Neutrons are expensive to produce The conversion rate is very low, so the time taken to build up usable amounts of U-233 are very long Reprocessing thorium fuel is less straightforward than with the uranium-plutonium fuel cycle The THOREX process has been demonstrated at small scale, but will require R&D to develop it to commercial readiness U-233 recycle is complicated by presence of ppm quantities of U-232 (radiologically significant for fuel fabrication operations at ppb levels) U-233 is weapons useable material with a low fissile mass and low spontaneous neutron source U-233 classified by IAEA in same category as High Enriched Uranium (HEU) with a Significant Quantity in terms of Safeguards defined as 8 kg compared with 32 kg for HEU

R&D requirements Fuel materials properties Fuel irradiation behaviour THOREX reprocessing Waste management /disposal U-233 fuel fabrication Systems development Scenario modelling Systems development Engineering design Materials Liquid salt chemistry and properties MSR fuel and fuel cycle Systems integration and assessment Safety

Fuel cycle scenario modelling Fuel cycle simulation computer programs are used to assess the impacts that different fuel cycle scenarios may have on: Uranium or Thorium ore requirements, Time and resources needed to create sufficient fertile material to start a Thorium ‘only’ reactor Ability to start a sustainable fast reactor fleet, Time at which feed of natural uranium is no longer required Packing density and inventory of a geological repository The practicalities of handling fresh nuclear fuel Processing of spent nuclear fuel Requirements for high level waste immobilisation technologies Fuel Cycle codes (e.g. NNL’s ORION code) are capable of assessing the impacts of the alternative U, Th based scenarios 15

Fuel cycle Building up a fleet with the aim of reducing dependency on U/Pu will take time. Reactor doubling time is an important consideration Some contention that alternative systems might give a different result But these underlying equations give confidence that the same limitations will apply to all workable systems

Relevance to thorium Long doubling times are relevant to: Initial build-up of U-233 inventory to get thorium fuel cycle to equilibrium For practical systems this timescale is very long and this will govern strategic analysis of transition to thorium fuel cycle using enriched uranium or plutonium/transuranic fuels Important for strategic assessments to account for impact of transition effects Subsequent expansion of thorium reactor fleet and rate at which thorium systems can expand to meet increasing demand This is a message I’ve been preaching for a long time, that transition effects in the fuel cycle are important.

Breeding ratio The breeding ratio (BR) is defined as: Mass of fissile material produced by fertile neutron captures Mass of fissile material consumed EXAMPLE: 1 GWth breeder reactor operating at 90% load factor would consume approximately 330 kg of fissile material per year – equivalent to 1 kg per full power day If a breeder reactor produces 1.3 kg of new fissile material by fertile captures per full power day, the breeding ratio is 1.3/1.0 = 1.3 and the breeding gain (BG) defined as BG = BR-1 is (1.3-1.0)/1.0 = +0.3 This is the standard textbook definition.

TD (full power days) = [MC + MO]/g Doubling time This is the time in which a breeder reactor would take to generate enough surplus fissile material to start off an identical reactor system The doubling time (TD) is the time needed to replace the total fissile inventory of the core MC (kg) plus the out of core fissile inventory MO (kg) For a system which consumes m kg of fissile material and has a net gain g kg of fissile material per full power day, the doubling time is: TD (full power days) = [MC + MO]/g = [MC + MO]/[(BR-1).m] = [MC + MO]/BG.m GOVERNING PARAMETERS: m is governed by the thermal power output only – 1 kg per full power day for 1 GWth output [MC + MO] and BG are dependent on the specific reactor design [MC + MO] typically a few thousand kg Large positive BG very difficult to achieve and 0.3 to 0.4 is about the highest claimed for any system This is again the standard textbook definition. TD (full power days) = [MC + MO]/BG.m The bottom line is that the numerator is several thousand kg and the denominator is 1 kg multiplied by BG < 0.4 Therefore doubling time is at best several thousand full power days and in practice longer still This applies generally to thermal and fast breeders and irrespective of the fissile nuclide (U-233, U-235 or Pu-239)

Application to MSR THERMAL SPECTRUM MSR FAST SPECTRUM MSR Based on simplistic scale-up of ORNL Molten Salt Reactor Experiment: 1.0 GWth; m = 1.0 kg/full power day; MC = 1500 kg U-233; MO = 3000 kg U-233; BG = +0.06 (estimated) TD = [MC + MO]/BG.m = (4500/0.06 x 1.0) = 75000 full power days (200 full power years) Probable scope for optimisation, but doubling time still likely to be very long FAST SPECTRUM MSR Based on Delpech/Merle-Lucotte et al TMSR-NM (non-moderated thorium molten salt reactor) core: 2.5 GWth; m = 2.5 kg/full power day; MC+MO = 5700 kg; BG = +0.12 TD = [MC + MO]/BG.m = (5700/0.12 x 2.5) = 19000 full power days (52 full power years) These are just illustrative examples. Thermal spectrum MSR has a reduced core mass, but then BG is lower as well and the result is just marginal breeding gain and very long doubling time. Fast spectrum MSR benefits from a larger breeding gain, but the core mass is higher. It is possible for a higher BG if you reprocess the salt more often (less fission products absorbing neutrons used to Breed), or add more Pu in place of U-233 into the core.

Hypothetical profile of installed capacity versus time for a breeder system Point A: Fast reactors first become available Segment AB: Fast reactors deployed using separated plutonium stock Point B: Initial deployment using available plutonium stocks ends Segment BC: Fast reactors deployed limited by doubling time Point C: Maximum capacity achieved Segment CD: Steady state operation in self sufficient mode Point D: Fast reactors converted to burners Segment DE: Fissile inventory drawn down to zero during planned phase-out ORION scenarios are necessarily tied to a specific set of reactor systems This leads to possible challenge that alternative systems might give a different result But the underlying equations give confidence that the same limitations will apply to all workable systems A kind of a 2nd Law of Thermodynamics for breeder reactors

Reactor parameters Parameter Value Units Unit size 1.6 GWe Initial core fissile loading 10.0 tHM Dwell time 4.0 years Recycle time 5.0 Net breeding gain (in breeding mode) +0.3 - Net breeding gain (in self-sufficient mode) 0.0 Net breeding gain (in burner mode) -0.40 Earliest fast reactor deployment 2040 Maximum fast reactor capacity 22.4

Generating capacity - transition from LWRs to fast reactors This assumes a Breeding Gain of 0.058 or 5.8% If you want to increase the energy output from your FR fleet you could use a different design FR with a higher BG; through either using higher density fuels (e.g. metal or nitride), employing radial reflector regions (more of a proliferation risk higher levels of fissile Pu in fuel), shorter dwell time (less Pu held up i.e. reduce the total dwell time through 5yrs cooling + 2yrs for repro and fabrication) 75 GWe target installed capacity FRs introduced at same rate as LWRs retire LWRs fuelled with UO2

Generating capacity - transition from LWRs to fast reactors The gap is plugged through reducing the total dwell time through 5yrs cooling + 2yrs for repro and fabrication) You need a pre-existing LWR programme of similar size to produce sufficient Pu to fuel Breeding in a FR programme. FR’s need 10Te Pu to start + 5Te Pu for core reloads – you can produce this in around 60 yrs lifetime of a LWR. In this case you can see how the stockpile of Pu from a LWR programme is able to build a FR programme which becomes self sufficient through breeding. It takes 75 GWe target installed capacity FRs introduced at same rate as LWRs retire FR Fuel dwell time is reduced

Generating capacity - transition from LWRs to fast reactors - Scenario (b) You need a pre-existing LWR programme of similar size to produce sufficient Pu to fuel Breeding in a FR programme. FR’s need 10Te Pu to start + 5Te Pu for core reloads – you can produce this in around 60 yrs lifetime of a LWR. In this case you can see how the stockpile of Pu from a LWR programme is able to build a FR programme which becomes self sufficient through breeding. Thorium reactor fleet power increases at a rate of 5GWe over 10 years assuming system as indicated in Doubling time example 10GWe of Th breeding FR’s introduced ~2040 FR breeders fuelled only with U-233 introduced ~2045

Conclusions Thorium is a valuable strategic alternative to uranium Sustainability remains one of the main drivers Radiotoxicity benefit is real, but modest Long term equilibrium radiotoxicity a simplistic measure Inherent proliferation resistance not proven for thorium Economics of thorium not known at present Minimum 15-20 year timeframe for commercial deployment (thermal systems) and longer timeframes for fast reactors Significant R&D programme required to progress technical maturity

Acknowledgements Kevin Hesketh Robbie Gregg Mike Thomas Chris Grove Richard Stainsby daniel.p.mathers@nnl.co.uk

Further information

Thorium history In the 1950s through to the 1980s, there were thorium research programmes for: Pressurised water reactors (PWR) Shippingport breeder core Germany-Brazil collaboration High temperature gas reactors (HTR) DRAGON (UK), Fort St Vrain (USA), Peach Bottom (USA), AVR (Germany) Molten salt reactors (MSR) Molten Salt Reactor Experiment (USA) The common driver for all these plants was to decouple nuclear expansion from uranium availability

Why did thorium research stall? Thorium cycle requires neutrons from uranium or plutonium fissions to get started U/Pu fuel cycle already established Large barrier to entry for a new system Technological issues THOREX reprocessing and fabrication of U-233 fuels

India/Lightbridge India Lightbridge Synergistic fuel cycle involving fast reactor and Advanced Heavy Water Reactors (AHWR) Fast reactor will breed U-233 in a thorium blanket U-233 will be recycled into AHWR fuel Lightbridge Seed/blanket assembly design for PWRs Low enriched uranium (LEU) seed region provides spare neutrons ThO2 blanket breeds U-233 Seed and blanket regions have different in-core dwell times

Pu/Th MOX AREVA are investigating PuO2/ThO2 MOX fuel for the eventual disposition of PWR MOX fuel assemblies PWR MOX fuel currently not reprocessed in France Held in long term storage pending eventual recycle in SFR fleet Requirement to cover all contingency that SFR fleet is not built Recycle of Pu from MOX fuel preferred over disposal PuO2/ThO2 MOX is presumed to be another option with potential advantage of low development cost and high stability as a final waste form Thor Energy undertaking PuO2/ThO2 MOX fuel qualification programme through a international consortium

Th-232 radiative capture cross-section

U-238 radiative capture cross-section

Decay heat and radiotoxicity Thorium-plutonium MOX fuel theoretically could be advantageous for UK plutonium disposition Detailed assessment by NNL of decay heat load and radiotoxicity per GWye shows there is only a marginal difference between Th-Pu MOX and U-Pu MOX This is a holistic calculation that accounts for the total decay heat outputs of different scenarios In the Th-Pu and U-Pu MOX cases, the decay heat is concentrated in the MOX assemblies, whereas in the UO2 reference case it is distributed over a larger number of UO2 assemblies

Core fissile inventory MC The fissile inventory of the core depends on a number of factors: Minimum critical mass for the system Thermal power output Specific rating of the core in MW/tonne Refuelling interval Refuelling strategy – single batch or multiple batch core loading KEY POINTS: The minimum critical mass can range over 3 orders of magnitude for different configurations (for example from 5 kg for a HEU research reactor core to several 1000 kg for a typical 1 GWe power plant) Workable designs typically nearer the upper end of the mass range and therefore MC is practically constrained to a few x 1000 kg Very important distinction between MC and m, which are orders of magnitude different for any practical system My main point is that engineering constraints force the core fissile mass to be at least 1000 kg. While it is possible to have a much smaller critical mass, if you want a large thermal output you need to spread the thermal load over a larger mass of fuel, otherwise the fuel will fail. The graph on the next slide confirms this for all current and projected commercial power reactors.

Illustration that large power reactors have a large fissile inventory based on survey or world reactors

Out of core fissile inventory MO For a conventional solid fuel reactor, this is the inventory in spent fuel awaiting reprocessing or being reprocessed, plus the inventory of fuel under fabrication, which depends on: Spent fuel cooling time tc Reprocessing time tr Fuel fabrication time tf For a fuel dwell time T, MO scales with MC: MO = MC x (tc+tr+tf)/ T For a liquid fuel system such as Molten Salt Reactor (MSR), there is an out-of-core inventory, which is the mass of fuel circulating through the heat exchangers This is the standard textbook definition. Out of core fissile inventory MO For a conventional solid fuel reactor, this is the inventory in spent fuel awaiting reprocessing or being reprocessed, plus the inventory of fuel under fabrication, which depends on: Spent fuel cooling time tc Reprocessing time tr Fuel fabrication time tf For a fuel dwell time T, MO scales with MC: MO = MC x (tc+tr+tf)/ T For a liquid fuel system such as Molten Salt Reactor (MSR), there is an out-of-core inventory, which is the mass of fuel circulating through the heat exchangers