Fake nuclear warheads exposed with neutron imaging

Princeton University researchers devise – and demonstrate – a tricky technique to verify nuclear arms without giving away any information about the weapons themselves. Angus Bezzina reports.

To join the nuclear club, a nation needs to have its nuclear warheads – such as this Titan nuclear intercontinental ballistic missile in an Arizona, US bunker – verified. But they want to protect their research and engineering from prying eyes, making verification difficult.
Michael Dunning / GETTY IMAGES

Does North Korean leader Kim Jong-un really have nuclear missiles at his disposal or is he all talk? A new way to prove he does – or not – could be around the corner.

US scientists have shown they can compare objects without revealing sensitive information, a method that could be used to verify nuclear warheads without exposing their structure or design.

This development has significant implications for global nuclear security, as the efforts of different nations to conceal inner workings of their warheads severely limits verification.

The work, published in the journal Nature Communications, was conducted by Sébastien Philippe from Princeton University in the US and colleagues.

It follows efforts of a team including co-authors Alexander Glaser and Robert Goldston which, in 2014, proposed an interactive protocol for warhead verification based on what are known as "zero-knowledge proofs".

Zero-knowledge proofs are a significant tool in modern cryptography. They show whether a claim is true or not without revealing any further information about a problem.

It sounds counterintuitive – how can you gain information without, well, revealing information?

“Suppose that there is a building separated into two identical rooms, each with a door opening outside,” says Willy Susilo, a computer scientist at the University of Wollongong in Australia and who was not involved with the study.

“From outside there does not appear to be a way to move between these two sides of the building without using the two external doors.

“[But] a person can prove to you that a secret passage exists between these two rooms without showing you anything else about it if they walk into one room of the building and exit from the other.”

Philippe and his colleagues came up with a physical zero-knowledge method to compare objects and tested it on a set of five-centimetre steel and aluminium cubes.

The cubes were arranged into different combinations designated either “true” or “false”. High-energy neutrons were fired through each arrangement to bubble neutron detectors on the other side.

These detectors were filled with superheated droplets of liquid spread throughout a clear gel. When a neutron struck one of these droplets, the droplet vaporised immediately, forming a bubble trapped in the gel.

These bubbles in the detectors told the researchers whether an arrangement was in a “true” configuration or not.

The superheated droplet detector on the left was exposed to neutrons, as indicated by the bubbles.
Alexander Glaser

To avoid revealing any additional information about the cubes, the gel detectors were preloaded with another set of specially designed bubbles.

If a “true” arrangement was presented, the sum of the preloaded bubbles and the signal detected through the cubes would equal the count produced by firing neutrons straight into the detectors, without any blocks between.

This way, the scientists found they could successfully perform comparisons of objects without gleaning specific data about the items being compared.

If developed for use in arms control, this method would collect bubbles from irradiation spat out from a suspected warhead alongside a preloaded set placed into the detectors by the warhead’s owner.

Should the number of the new and preloaded bubbles equal the count produced by firing neutrons into the detectors with nothing in front of them, the weapon’s nuclear threat could be confirmed real. Otherwise, it would be exposed as a fake.

Philippe and his colleagues’ method does not require classified data be exposed, nor does it use electronic components that might be susceptible to tampering or cheating.

Susilo reiterates the importance of this: “Wherever we use zero-knowledge proofs there is always a possibility of cheating.

“It is critical to minimise the risk of this as much as possible.”

The researchers are focusing on refining their method to suit actual nuclear warheads and be accepted by leaders from around the world.

Contrib angusbezzina.jpg?ixlib=rails 2.1
Angus Bezzina is a writer from Sydney, Australia.
Latest Stories
MoreMore Articles