Tritium Production Pathways
Tritium can be made by neutron irradiation in fission reactors or bred in a fusion blanket; the two routes differ in feedstock, by-products, and infrastructure.
Two routes to the same isotope
The established route irradiates lithium-bearing targets in a fission reactor: reactor neutrons convert lithium-6 to tritium, and targets are later processed to extract the gas. The fusion route breeds tritium in a lithium blanket around a deuterium-tritium plasma using the device's own 14 MeV neutrons, as studied for the breeder (Hyperion).
What differs between them
- Feedstock: both use lithium-6, but the neutron source differs (fission core vs fusion plasma).
- Co-products: the fusion route co-produces helium-3 and 14 MeV neutron services from one platform.
- Coupling: fission-route output is tied to reactor operation; fusion-route output is tied to plasma operation.
- Maturity: reactor irradiation is demonstrated; fusion breeding is at design and simulation stage.
Neither route is presented here as superior in the abstract. The relevant point for supply security is diversity: a domestic fusion pathway would add a route that does not depend on the same facilities or feed streams as the existing one, reducing single-point exposure.
Diversity is the strategic goal
The comparison is not a contest to crown one route but a case for holding more than one. A supply that depends on a single production method, however mature, fails entirely when that method is interrupted. Two independent routes that do not share facilities, feedstock preparation, or processing lines mean an interruption in one still leaves make-up flowing from the other. For a decaying isotope, where any pause erodes inventory, that independence is the difference between a manageable disruption and a falling stockpile, which is the strategic reason to develop a fusion pathway alongside the established one.
Because the breeder is not yet built, this comparison is qualitative and forward-looking; capacity figures for the fusion route are computed design targets, not measured throughput.