Question #6

Tritium is the main fuel ingredient in the plasma of MFE and IFE systems based on the D-T fuel cycle. Since tritium is not a naturally existing isotope, attaining tritium self-sufficiency is necessary for self-sustaining fusion plants operating on the D-T fuel cycle. Tritium is bred in a lithium-containing blanket surrounding the plasma. Once tritium is generated, it needs to be collected, processed, and redirected to the plasma. The tritium fuel cycle involves many subsystems whose physical and operational characteristics will dictate the success in achieving tritium self-sufficiency.

The achievable tritium breeding ratio is predicted by performing neutronics calculations. It depends on the type of the breeding material as well as the coolant and structural materials used in the FW/blanket subsystem. The calculated achievable TBR should account for the 3-D geometrical configuration of the chamber including penetrations. Hence, in addition to the dependence on the blanket type, the achievable TBR might depend on the plasma confinement concept considered. Moreover, the geometrical and spectral differences of the neutron source in MFE and IFE chambers affect the achievable TBR. The calculated achievable TBR for a given FW/blanket concept is uncertain due to the uncertainty associated with system definition and the inaccuracies in predicting the TBR. The latter includes the uncertainty associated with the geometrical modeling, calculational methods, and basic nuclear data.

The required TBR in a fusion system must exceed unity by a margin that accounts for calculational uncertainties, tritium losses and radioactive decay during the period between production and use, tritium inventory in the plant components, and supplying inventory for startup of other fusion plants. To accurately determine the required TBR, one has to consider the entire fuel cycle for the D-T plant. The tritium fuel cycle involves many subsystems. The main subsystems are plasma exhaust and vacuum pumping, FW/blanket, plasma-facing components (PFC), fuel clean-up, isotope separation, fuel management, storage, and fueling. Simulation of this cycle, including the dynamic behaviour is required for accurate evaluation of tritium build-up and consumption/losses in this closed cycle.

The required tritium breeding margin depends on many system parameters. These include the desired doubling time, tritium inventory in the different components, and the tritium extraction and processing system utilized. In addition, tritium fractional burn-up in the plasma impacts the tritium inventory in some of the components, such as the plasma fueling system, and hence affects the required TBR. The required TBR, therefore, depends on the plasma confinement concept used, the breeding blanket concept, and the tritium extraction and processing system. The required TBR is also uncertain due to the uncertainty involved in the performance characteristics of the plasma and other subsystems of the cycle.

To attain tritium self-sufficiency, the calculated achievable TBR must exceed the required TBR. Uncertainties in predicting both the achievable and required TBR should be addressed. In this report, the potential of the plasma confinement concepts in both MFE and IFE and the breeding blanket concepts for attaining tritium self-sufficiency will be assessed. Possible plasma and technology R&D required to reduce the uncertainties and increase the potential for achieving tritium self-sufficiency will to be identified.

The first generation of fusion ignition machines are designed without tritium breeding blankets and rely on the available tritium resources for supplying their fuel. These resources are decreasing due to radioactive decay and reduced production rate. An important issue, to be addressed in this report, is whether there is a time window for the availability of tritium to supply the tritium requirements for the ignition machines. This time window will impact the schedule for developing tritium producing chamber technologies.

The breeding potential varies substantially with FW/blanket concepts. The achievable local TBR depends on the Li bearing breeder type, Li enrichment, coolant, and structural material. Neutron multipliers, particularly beryllium, could enhance the tritium breeding potential of almost all breeders.

The inherent breeding capacity of numerous breeders considered in previous fusion designs over the past 30 years is illustrated in Fig.1. A fairly thick breeding zone without structure or neutron multiplier is considered for each breeding material. Lithium enrichment does not always help the breeding. Breeders with natural Li provide the highest TBR except for LiPb and LiSn. The TBR could optimize at higher enrichment when structural materials and multipliers are included in the blanket. In realistic designs, the structure, configuration, and penetrations will degrade the achievable overall TBR compared to the values shown in Fig. 1.

The breeders could be divided into three groups according to their breeding potential. The first group includes liquid Li and LiPb which have the largest breeding potential with local (full coverage) TBR values greater than 1.7 without structure or multiplier. The second group contains Li2O, Flibe, and LiSn. These breeders have medium breeding potential with local TBR values in the range between 1.2 and 1.4 without structure or multiplier. To achieve tritium self-sufficiency with these breeders, the structure content needs to be minimized and/or moderate amount of neutron multiplier should be added. The third group includes several ceramic solid breeders, such as Li2ZrO3, Li2TiO3, Li4SiO4, and LiAlO2, which have poor breeding potential and need substantial amount of neutron multiplier to achieve adequate breeding.

Using structural material in the FW/blanket results in degrading the achievable TBR. The extent of degradation depends on the structural and breeding materials used. The amount of structure in the FW and front layer of the blanket has a much more severe impact on tritium breeding relative to the structural content in the bulk of the blanket. Depending on the breeder material and structure type and content, up to 20% degradation in TBR might result. Candidate structural materials used in conceptual designs include austenitic steel, ferritic steel, vanadium alloys, SiC/SiC composites, and C/C composites. Because of their ability to handle high surface heat fluxes and operate at high temperatures, refractory alloys, such as W, Ta, Mo, and Nb alloys, have also been considered.

For liquid Li breeder, vanadium structure results in the least impact on tritium breeding. SiC structure slightly enhances tritium breeding in a LiPb system. W and V alloys yield the smallest degradation in TBR when used as structural material with Flibe of LiSn breeders. For solid breeders, using vanadium structure gives the least impact on tritium breeding. Except for LiPb and LiSn, some of the TBR degradation resulting from the structure can be recovered by enriching the Li in 6Li. However, the net effect is still a reduced TBR compared to the case without structure. Eliminating the structure completely as in concepts used in IFE chambers and liquid wall concepts proposed in MFE designs will have an advantage regarding the tritium breeding capability.

While the liquid breeder can serve also as the coolant in a self-cooled concept, separate coolants, such as water or helium gas, could be utilized with solid and liquid breeders. Due to its low density, He gas has negligible impact on tritium breeding. On the other hand, the large neutron moderation in water helps enhancing tritium breeding from 6Li. However, the large absorption tends to decrease tritium breeding. The net effect depends on the breeding and structural materials used. In typical liquid and solid breeder designs, using 20% water coolant in the FW/blanket system reduces the TBR by ?? and ??, respectively.

Different neutron multipliers can be used to enhance the achievable TBR. The enhancement depends on the breeding and structural materials used. Beryllium is the best neutron multiplier followed by Pb, Be2C, and BeO. No multiplier is needed in concepts utilizing Li or LiPb as a breeder. If Flibe and LiSn are used without structure, adequate breeding might be achievable without a multiplier. However, it seems to be necessary to add a neutron multiplier when Flibe, LiSn, or solid breeders are used in a FW/blanket concept with structure.

There are several geometrical, spectral, and temporal differences between the IFE and MFE systems that could impact the achievable TBR. While a cylindrical or toroidal chamber surrounds a volumetric distributed source in MFE systems, a nearly spherical chamber in IFE plants surrounds a point neutron source. As a result, source neutrons in IFE chambers impinge on the FW/blanket in a more perpendicular direction than in MFE chambers. This leads to lower tritium production rate at the front with lower radial gradient. The achievable TBR will be similar in the two systems if the blanket is quite thick. On the other hand, for relatively thin blankets utilizing the same materials, the local TBR is expected to be lower in IFE chambers. However, since the chamber size is decoupled from the size of the driver in IFE plants, thicker blankets are easier to accommodate in IFE chambers than in MFE chambers.

Fusion neutron interactions in the highly compressed target results in considerable softening of the neutron spectrum incident on the FW/blanket in IFE chambers. These neutrons can have average energies as low as 10 MeV depending on the target fuel compression ( ρR). This tends to decrease the achievable local TBR The combined geometrical and temporal effects result in local TBR values in IFE chambers that are lower than those in MFE chambers by up to 5% depending on the FW/blanket concept used.

While steady state or long pulse operation is envisioned for MFE plants, neutrons are born over 10-100 ps pulses in IFE plants. The neutrons emanating from the target traverse the chamber in 40-100 ns depending on the radius. The of flight spread of source neutrons with the softened spectrum along with the neutron slowing down in the blanket results in tritium production being spread over several hundreds of nano seconds following each pulse. This temporal effect does not affect the time integrated local TBR. However, it could affect tritium permeation and extraction.

The calculated achievable TBR should account for the 3-D geometrical effects. The plasma chamber configuration differs with the confinement concept and, therefore, impacts the blanket coverage and the achievable TBR. Confinement schemes, like tokamaks, STs, and stellarators, in which the plasma is linked by TF coils, require a divertor (or limiter) system that faces the plasma. The overriding design consideration for the divertor will be power exhaust (accepting ~10 MW/m2 of power continuously and reliably) and particle control (pumping fuel and helium ash out of the plasma chamber). Hence, tritium breeding will be compromised or absent in the divertor region. Confinement systems, like FRCs and spheromaks, in which there are no TF coils linking the plasma, have a potential advantage in that power and particles can be diverted along field lines that leave the plasma chamber. Hence, the power and particle control systems need not compete with the tritium breeding systems for space facing the plasma in these confinement systems. No divertors or limiters are needed in an IFE system.

Conventional tokamaks rely heavily on the inboard (IB) and outboard (OB) blankets to breed all the tritium needed for plasma operation. The contribution to the overall TBR from a blanket installed behind the divertor system is insignificant due to the large attenuation of neutrons by the sizable divertor structure. A tokamak with a single null divertor will have about 5% higher breeding capability than a double null design that uses the same FW/blanket concept. Due to space limitation in the IB side of a tokamak, usually a thinner blanket is used in the IB region with a smaller local TBR than the thicker OB blanket. This tends to have a negative impact on the achievable overall TBR.

In spherical tokamaks, it is unlikely that a breeding blanket could be installed in the space-constrained IB side. Therefore, spherical tokamaks depend entirely on the OB blanket for tritium breeding. However, this is not expected to drastically affect the overall TBR since the low aspect ratio results in less than 10% IB coverage. As the aspect ratio of a ST increases, the OB blanket coverage fraction decreases and it becomes difficult for the OB blanket to provide the required tritium. Hence, STs with lower aspect ratios will have higher breeding capability.

In stellarators, the aspect ratio is high and a blanket with uniform thickness surrounds the entire plasma. Although the coverage fraction of the divertor in the stellarator is larger than in the tokamak, thin divertor plates or baffles are used allowing for substantial breeding in the divertor region. In linear confinement concepts such as the FRCs and tandem mirrors, elongating the cylindrical chamber can reduce the end losses. This allows increasing the blanket coverage which helps enhancing the overall breeding potential. The linear confinement concepts also allow for using uniformly thick blankets. In IFE plants, the chamber geometrical configuration allows for full blanket coverage with blankets that could be as thick as needed at all locations in the chamber. This allows for achieving an overall TBR very close to the local value.

Penetrations are required in MFE chambers to accommodate heating and current drive systems, and diagnostic systems. Heating and current drive systems will probably be required for fusion power plants based on any magnetic confinement scheme. Given our present ignorance regarding the heating and current drive requirements for the various alternate confinement schemes, it is not possible to reach any conclusion regarding relative advantages in tritium breeding of any one magnetic confinement scheme over the others. In particular, the predicted lower non-inductive current drive requirements of STs relative to advanced tokamaks results mainly from the greater optimism of the proponents of STs regarding our ability to control the pressure profile and, thereby, the profile of the bootstrap current. This cannot reasonably be translated into a tritium breeding advantage for STs.

Stellarators will require heating systems, while compact stellarators (which carry a toroidal current) may also require non-inductive current drive for current profile control. What advantage stellarators may achieve in tritium breeding from reduced (or absent) current drive systems will probably be balanced by the need to keep the tritium breeding blanket as thin as possible to insure that the external coils producing the rotational transform of the magnetic field are as close to the plasma as practicable. FRCs with rotating magnetic field current drive would appear to require substantial antenna systems, which would compete with tritium breeding systems for space facing the plasma.

The penetrations for heating and current drive are normally placed in the OB side of the toroidal concepts. This is the region where the local TBR is highest. As a result the impact on reducing the achievable overall TBR will be greater than predicted by the fraction of FW area occupied by the penetrations. In previous conceptual MFE power plant designs, the area taken by the heating and current drive penetrations amounts to 1-2% of the FW area and the net effect on the overall TBR is about 2-3% reduction.

IFE power plants have a clear advantage regarding the impact of penetrations on the achievable TBR. No heating or current drive systems are needed in IFE plants. The penetrations in an IFE chamber provided for the laser or ion beam fusion driver represent less than 0.5% of the FW area in direct drive concepts with up to ~100 beam ports. In indirect drive concepts, the fraction taken by the beam ports is much lower. Hence, the impact of chamber penetrations on the achievable overall TBR is minimal in IFE plants.

In commercial fusion power plants, maintenance schemes utilize ports that do not interfer with the breeding blanket. Maintenance is achieved through doors or ports in the vacuum vessel behind the breeding blanket. Hence, the maintenance systems do not impact the achievable TBR. Some diagnostics will be required for any confinement concept. They will be particularly important for concepts, like advanced tokamaks, STs, and compact stellarators, in which optimized performance is to be achieved through control of plasma profiles. The diagnostic penetrations are usually much smaller than those required for heating and current drive. Therefore, they are not expected to have a dominant impact on the TBR.

To guide the design process of a specific blanket concept, a series of parametric 1-D analyses is usually established at an early stage of the design. The 1-D model used in the calculations depends on the confinement concept. A toroidal cylindrical geometry modeled around the machine axis is more appropriate for toroidal machines such as tokamaks and STs. In this model, the IB and OB regions are modeled simultaneously to properly account for the toroidal effects. The cylindrical model around the plasma axis is more suitable for the large aspect ratio stellarators and linear machine such as the FRC. This model is used also in toroidal machines to determine the neutronics parameters at the top and bottom (divertor region). For IFE chambers, a spherical geometry is used with a point source at the center. The point source emmits neutrons and gamma rays with spectra determined by performing a separate target neutronics calculation.

In the early stages of the design when several iterations are needed, the overall TBR is estimated by coupling the 1-D local TBR values obtained in the different regions surrounding the plasma with the appropriate coverage fraction. It needs to be emphasized that the nuclear coverage fraction (NCF) should be used rather than the FW area fraction. The NFC is defined as the fraction of the source neutrons incident directly on the specific region. In toroidal facilities, the NCF of the OB and IB regions depend on the aspect ratio. Increasing the aspect ratio results in decreasing the OB NCF and increasing the IB NCF. When estimating the overall TBR, provisions should be made for the elements that degrade the breeding, such as penetrations, assembly gaps, and side walls. In addition, accurate modeling of the layered heterogeneous configuration of the breeding blanket in the 1-D model is essential for accurate estimation of the TBR. Some error will be introduced if homogenized zones are used.

Upon converging on a reference design, a detailed 3-D model for the reference configuration is necessary to confirm the achievable TBR. Past experience with ITER, commercial MFE plants (ARIES), and commercial IFE designs (HIBALL and LIBRA) indicate that the overall achievable TBR estimated from the 1-D calculations coupled with coverage fractions are within 3% from the value obtained from detailed 3-D calculations. Assuming that a detailed 3-D calculation is used to determine the achievable overall TBR, uncertainties in the TBR resulting from possible modeling approximations such as zone homogenization are estimated to be ~1-2%.

The tritium fractional burn-up in the plasma impacts the tritium inventory in some of the components, such as the plasma fueling system, and hence affects the required TBR. Increasing tritium burn-up reduces the required TBR and, hence, improves the potential for achieving tritium self-sufficiency.

A zero-D analysis shows that there are two general means of increasing the tritium burn-up fraction. One option is to operate with a tritium-lean fuel mixture, which increases the required confinement capability. The other approach is to improve the particle exhaust. Given the fact that confinement capability strongly drives the unit size and cost, it is implausible that increasing the tritium burn-up fraction through increased confinement capability can provide much net gain in power plant attractiveness for confinement concepts, like tokamaks, which are relatively well understood. Improved helium pumping efficiency (as measured by tau_He*/tau_E) can improve the tritium burn-up fraction, as can helium enrichment (relative to the DT fuel) in the particle exhaust stream (as measured by tau_He*/tau_DT*). Any advantage assigned to a particular magnetic scheme for improved particle exhaust must be based on some qualitative difference in power and particle exhaust system. Schemes, like FRCs, spheromaks, and dipoles may have an advantage in this respect.

b. The cryogenic pump: The typical regeneration time for the cryopump is about 30 minutes. The tritium flow rate to the cryopump is less than 600 g/hr (3000 MW fusion power, with plasma burn-up fraction of 3%.). Therefore, the tritium inventory on the cryopump is less than 300 g.

Regarding the chamber technology concepts, it is clear that concepts utilizing liquid lithium or LiPb as breeder have the largest potential for achieving tritium self-sufficiency even if a large amount of structural material is used. LiSn and Flibe can achieve tritium self-sufficiency without a need for a separate multiplier if utilized in designs with very small amount of structure. Eliminating the structure completely as in IFE systems or MFE systems with liquid walls enhances the potential of achieving tritium self-sufficiency with these breeders. Among solid breeder candidates, Li2O has the best chance for achieving tritium self-sufficiency. However, the structural material and coolant required in solid breeder concepts imply that a neutron multiplier should also be used to achieve tritium self-sufficiency. The does not impact the required TBR. The FW/blanket concept impacts the required TBR through the effect of material choice on the tritium inventory. However, tritium inventory in the FW/blanket system represents a very small fraction of the total tritium inventory in the plant. Hence, the required TBR is, in general, independent of the chamber technology concept utilized.

For the plasma confinement concept to have a large potential for attaining tritium self-sufficiency, it needs to have high tritium burn-up fraction in plasma and allow for large coverage with the breeding blanket. While a large tritium burn-up fraction (<30%) can easily be achieved in the IFE targets, it seems to be more difficult to increase the tritium burn-up fraction to that level in several MFE systems. However, there is no significant reduction in the required TBR if the tritium burn-up fraction is increased above ~10%. Since all MFE and IFE confinement concepts are expected to achieve values above 10%, the tritium burn-up fraction is not the dominant factor in determining the confinement concept’s potential for achieving tritium self-sufficiency.

The ability of a confinement concept to allow for large breeding blanket coverage is determined by the chamber geometrical configuration and amount of penetrations required. The IFE systems have a clear advantage since no divertors, limiters, or heating and current drive systems are employed. In addition, beamline penetrations are very small covering less than 0.5% of the FW. Blankets can be made as thick as needed in IFE chambers without impacting the high cost driver. The large blanket coverage in IFE chambers allows using breeding materials that have attractive thermal-hydraulic features but marginal breeding such as Flibe and LiSn. In addition, due to the lack of magnetic fields in IFE chambers, it is also easy to employ flowing thick liquid breeder concepts without structure. Among the different MFE concepts, FRCs and spheromaks do not require divertors or limiters. Other concepts such as stellarators, STs, and tokamaks require divertors or limiters resulting in limiting the overall TBR. There is no clear advantage for any of the MFE confinement concepts with respect to the number and size of chamber penetrations required. It should be emphasized that, even though some MFE confinement concepts suffer from reduced blanket coverage, tritium self-sufficiency can still be achieved in such concepts with careful design utilizing blanket concepts that have high breeding potential.

The largest sources of uncertainty in predicting the achievable TBR are due to uncertainties in the basic nuclear data and the geometrical approximations associated with the calculational model. This amounts to a total of about 7-9%. As a result, the FW/blanket in a power plant is designed for an overall TBR that exceeds unity not only by the calculation uncertainty in addition to the tritium breeding margin that depends on the operating parameters and tritium processing system. For example, if the required breeding margin is 0.03, the design should allow for an overall TBR of 1.12. This implies that the actual TBR realized in the plant, which can not be verified until the plant starts operation, could be in the range between 1.03 and 1.21 for a +/-9% uncertainty. Underbreeding (TBR < 1.12 for the above example) will place the plant operation at risk as the tritium bred may not suffice for machine operation. It is, therefore, imperative to conservatively design the power plant with overbreeding (TBR >1.12 for the above example) providing that design solutions are established for reducing the breeding level if needed. The adjustment in the design could take place during the second period of operation after changing out the FW/blanket and realizing higher breeding than predicted.

If the life time of the blanket is 2 years, the total tritium produced in the blanket is 370 Kg. If the accuracy of the tritium production rate is 10 %, the error of the tritium production rate during the two years blanket life is 37 Kg. Therefore, the error due to the tritium breeding calculation far exceed the total tritium inventory and tritium loss rate in the power plant.

Based on comments from Drs Filatov and J. Anderson, there will be no extra tritium supply to ITER, from either RF and US defense program. Therefore, the only tritium supply available to the ITER type device will be from Canada. The tritium production rate from CANDU reactors is about 2.5 Kg/y, assuming all the CANDU reactors are in operation. Therefore, the availability of tritium supply from Canada for ITER operation strongly depends on the lifetime of the CANDU reactors, as well as the starting time of the ITER device. Early shut down of the CANDU reactors, and /or delay starting of the ITER, will waste some the tritium produce by CANDU due to the decay.

If the lifetime of the CANDU s are 40 years, and if ITER starts about 10 years from now, and assuming the parameters of ITER remains to be the same as the ITER-EDA device, there will be sufficient tritium supply to ITER from Canada. However, if the CANDU lifetime is only 30 years, or the ITER starting operating day is much further away than 10 years, the tritium supply from Canada to ITER will be very marginal.

A key question is where will we obtained tritium for post ITER operation. There will be either an extended phase of ITER, or there will be another fusion device following ITER. In either case, all the tritium supply from CANDU will be exhausted, and no sufficient tritium will be available for the starting tritium supply for the next device. Therefore, we can not only assess the fuel supply for ITER along. We also need to assess the fuel supply for the next device. For this reason, even there is sufficient tritium supply for the ITER like device from Canada, we may still have to breed tritium.

Snowmass Hot Topic

Chamber Science and Technology

Key Question number 6

Potential for Achieving Tritium Self-Sufficiency