Antineutrino detectors could expose covert plutonium production in fusion reactors

A fusion plant that quietly starts making plutonium would not have to be torn open to be caught. A new paper in Phys. Rev. Applied says a relatively small onsite antineutrino detector could confirm the production of a few kilograms of plutonium over 30 days, giving safeguards a way to watch for covert fissile-material work without touching the machine.

The logic is straightforward: if an operator introduces fertile material such as uranium-238 into a region exposed to the reactor’s intense neutron flux, that material can breed fissile isotopes. The authors, Alexander Glaser, Robert J. Goldston and Patrick Huber, argue that antineutrinos provide a clean signature of that process because they pass through shielding so easily. In their view, the detector would sit close enough to the system to sample the antineutrino emission tied to the reactor’s operating state, while normal power production would continue with near-zero interference.

That is what makes the safeguard concept attractive and also what makes it hard. The paper explicitly tested the signal against cosmogenic background and antineutrinos produced by neutron activation of reactor components, two sources that can blur the picture around a working fusion system. Even so, the result points to a detector that could flag suspicious plutonium buildup on a timescale that matters: long enough to separate routine operation from a hidden breeding campaign, short enough to matter before material accumulation becomes large.

The concern is not new. In 2012, Glaser and Goldston laid out three fusion proliferation pathways: clandestine production in an undeclared facility, covert production in a declared facility, and breakout at a declared facility. Those scenarios are now more relevant as fusion moves from lab physics toward commercial hardware. The Fusion Industry Association said in July 2024 that the sector had drawn more than $7.1 billion in total investment, with 45 private fusion companies in its survey and more than $900 million in new funds added over the previous year.

Antineutrino monitoring has been building a safeguards pedigree for years. A 2001 Sandia and Stanford paper argued that cubic-meter-sized detectors could monitor reactors non-intrusively and automatically. In 2019, Christopher Stewart, Abdalla Abou-Jaoude and Anna Erickson showed in Nature Communications that antineutrino monitors could help detect diversion scenarios in both light-water and sodium-cooled reactors. That work fits alongside the International Atomic Energy Agency’s broader safeguards system, which its 2024 Safeguards Statement said was applied in 190 states.

For fusion, the attraction is obvious: a detector that can watch for forbidden plutonium production while the machine runs and the operators keep their hands off it. The problem is just as clear: the signal is still filtered through reactor noise, background radiation and modeling assumptions, so the idea looks promising as a safeguard layer, not yet as a settled tool.