New types of imager could help spot smuggled nuclear materials


Much as a smoke detector gives only a vague idea of where a fire is, current methods to detect smuggled nuclear materials are slow and imprecise. But a new technique that images nuclear materials based on the neutrons and gamma rays they shed can locate these dangers in record time, scientists report.

“It’s an elegant method,” says Alexander Glaser, a physicist at Princeton University who works on nuclear weapons verification and was not involved with the new study. If it proves itself in real-world scenarios, he says, the new approach could strengthen border security and help map radioactive contamination at disaster sites like Chernobyl and Fukushima.

Nuclear power plants and weapons centers have long kept a close eye on fissile materials—the stuff of atomic bombs—using radiation portal monitors (RPMs) to screen people or vehicles leaving a site. Resembling airport metal detectors, RPMs detect neutrons and gamma rays emitted by radioactive substances. They can distinguish the low-level radiation from everyday materials such as bananas or cat litter from the stronger signals from materials that actually pose a danger like plutonium or highly enriched uranium.

Security experts became more concerned about the proliferation of nuclear materials after the collapse of the Soviet Union in 1991 and the 9/11 attacks in 2001. RPMs installed worldwide between 1993 and 2019 flagged 290 confirmed or likely incidents of nuclear trafficking. But if inspectors are rushing to find and defuse a bomb in a shipping container, it would help to know precisely where the device is located. That’s where the new approach comes in.

Bo Cederwall, a nuclear physicist at the KTH Royal Institute of Technology, had the idea while working at France’s National Large Accelerator for Heavy Ions. Those experiments involve blasting atomic nuclei with a particle beam that knocks off neutrons, leaving energetically excited nuclei that radiate gamma rays. The scientists measure the timing and energies of the gamma rays and neutrons, which serve as fingerprints to distinguish one nucleus from another, enabling researchers to sift out the rarest nuclei for further study.

A few years ago, Cederwall realized that such an approach, coupled with machine learning methods, might come in handy for zeroing in on plutonium and other radioisotopes that, when they decay, also emit gamma rays and neutrons. “I saw a chance to bring new ideas and fresh blood into the game,” he says.

Since 2017, Cederwall has worked with Swedish authorities on nuclear safeguards and security technologies. Now, they are assessing spent fuel at a former research reactor site with thousands of drums filled with radioactive waste. “I wanted to help them figure out what the heck is inside those drums” without cracking open the lids, Cederwall says.

The new technique relies on detectors that emit light when struck by either a neutron or a gamma ray and measure the time of arrival with nanosecond precision. Suppose two detectors sit face to face, separated by 1 meter or so, and that a nucleus decays and emits a neutron that hits one detector and a gamma ray that hits the other. The difference in the arrival times, when accounting for the detailed physics of the nuclear decay process, defines a fuzzy, somewhat spherical shell in space in which the nucleus could have been. Timing many neutron–gamma ray pairs with several detectors produces a set of probability shells that should intersect at a point—the location of the source.

As a proof of principle, Cederwall and colleagues focused on detection of californium-252, a readily available radioisotope widely used as a proxy for weapons-grade plutonium. Their prototype neutron-gamma emission tomography (NGET) detector looks a little like two sets of four magnum wine bottles installed on either side of an aluminum RPM-like frame. Analyzing scores of collisions in a matter of seconds, the researchers found they could quickly pinpoint the source to within 4 centimeters of its actual location, as they report today in Science Advances. Some modest tweaks should shave that to about 1 centimeter, Cederwall says.

The ability to pinpoint a source may offer a “paradigm shift” in nuclear safeguards, Cederwall asserts. NGET detectors might also be shrunk to fit on a drone. That offers “a really fascinating possibility” of quickly mapping radiological contamination at disaster sites or environmental surveying, he says.

“I don’t think that’s a far-fetched claim,” says Brian Quiter, a nuclear physicist at Lawrence Berkeley National Laboratory (LBNL), whose team is also working on a drone-based nuclear materials detector. But NGET must still prove itself in real-world scenarios, he says.

One big challenge is that the real world is messy: Smuggled nuclear materials could be cocooned in materials that deflect neutrons streaming from the source. The californium in the Swedish team’s experiments was “not sitting in a container full of stuff,” notes Kai Vetter, a nuclear physicist at LBNL.

Tuning out scattered neutrons is a challenge, Cederwall acknowledges. But NGET’s probabilistic machine learning approach, he argues, should make it “less sensitive to scattering than other techniques.”, 19 May 2021