How to Weigh Plutonium From Across the Room

To know how much plutonium sits inside a sealed container—and whether it's the kind that fuels a reactor or the kind that arms a warhead—you don't open it. You read the light coming off it.

Here's why that works, why it was quietly failing, and how a thermometer balanced on the edge of physics fixed it.

Radioactive elements have fingerprints made of light

An unstable atomic nucleus decays. When it does, it throws off gamma rays—photons (particles of light) at the very high-energy end of the spectrum, far more energetic than visible light. The key fact: each radioactive element emits gamma rays at specific, characteristic energies. Plutonium's set of energies is not uranium's. Match the pattern of energies streaming out of a container against a known catalogue, and you know what's inside without touching it.

This is the backbone of nuclear safeguards—the international accounting system that checks whether the material at power plants and weapons sites matches what a country has declared, and flags whether any has gone missing.

The problem: X-rays wearing the gamma ray's disguise

The same atoms also emit X-rays—lower-energy photons produced when the atom's electrons rearrange themselves, a separate process from nuclear decay. And some of those X-rays land at almost exactly the same energies as the gamma rays you're trying to read.

Picture trying to pick out one voice in a room where a second person is talking at the same pitch. The gamma-ray "voice" carries the information you want. The X-ray "voice" sits right on top of it. If you can't separate them, your count of how much material is present goes fuzzy—and fuzzy is exactly what you cannot afford when the question is "reactor fuel or weapon?"

You can't remove the X-rays. But if you knew their exact energies with high precision, you could subtract them—model the interfering voice and take it out of the recording. That subtraction is only as good as your knowledge of where the X-rays sit. And that knowledge was, until recently, too coarse to do the job cleanly.

The fix: a thermometer that reads photons one at a time

To pin down the X-ray energies, NIST used transition edge sensors. The physics is worth understanding, because the same trick generalises.

A superconductor is a material that, cooled below a certain temperature, conducts electricity with zero resistance. Warm it past a sharp threshold and it snaps back to being an ordinary metal with normal resistance. That threshold is knife-thin: a tiny temperature rise flips it from zero resistance to measurable resistance.

The sensor is a thin superconducting film held right at that edge, a fraction of a degree above absolute zero (the coldest temperature physically possible, about –273 °C). A photon strikes the film and deposits a minuscule pulse of heat. Because the film is balanced on the transition, that tiny warmth produces a sharp, measurable jump in resistance.

Now the payoff. The size of the jump is proportional to the photon's energy: a more energetic photon deposits more heat and pushes the resistance further. So the sensor doesn't merely detect photons—it measures the energy of each one individually, with far finer resolution than conventional detectors, which lump nearby energies together into a blur. That resolution is the whole point. To subtract two signals sitting almost on top of each other, you must resolve them apart in the first place.

Using an array of about 250 of these sensors, the team measured the X-ray energies of uranium, plutonium, and neptunium in the exact range where they overlap with gamma rays, cutting the uncertainty by a factor of three to eight over previous measurements. Better numbers in the catalogue mean a cleaner subtraction for everyone reading gamma-ray fingerprints afterward.

Why the sharper number decides everything: isotopes

The stakes come down to isotopes—versions of an element with the same number of protons but a different number of neutrons. They are chemically almost identical but behave very differently in a nucleus.

Take uranium. In nature only 0.7% of it is uranium-235, the isotope that sustains a fission chain reaction. Enrich it to a few percent U-235 and you have reactor fuel. Enrich to around 90% and you have a weapon. Same element. The isotope ratio is the entire difference between a power plant and a bomb.

You read that ratio from the gamma-ray pattern—but only if the pattern is clean. Once the interfering X-rays can be subtracted precisely, the ratios sharpen, and you can tell fuel from weapons material more accurately and faster. Speed is not a side benefit: nuclear material gets reassessed at each processing step, and every hour spent waiting on a measurement is an hour it sits idle and unaccounted for. Faster reads mean lower cost and less material stuck in limbo.

The portable idea: put your detector on a threshold

Here is the move worth carrying away, because it is more general than nuclear physics.

To detect something faint, the brute-force instinct is to build a bigger, more powerful sensor. The smarter route: position your instrument at a threshold—a point where a system is poised to flip between two states—and let the faint signal push it across. Near that edge, a tiny input produces a large, readable output. You are borrowing the system's own instability as a free amplifier. The transition edge sensor works precisely because the superconductor is sitting where the smallest heat pulse causes a disproportionately large resistance change.

The reason this works, stated plainly: near a threshold, the system's response to input is steep—a small nudge in energy maps to a large swing in output. Away from the threshold, the same nudge barely registers. Sensitivity is the steepness of that response curve, and it is steepest at the edge.

The same steepness explains why a neuron, held just below its firing voltage, can be tipped into a full spike by a single small input—the cell is engineered to sit near its threshold precisely so it stays sensitive. It explains why a supersaturated solution, poised at the edge of crystallising, solidifies the instant one speck disturbs it. In each case the sensitivity comes not from the size of the detector but from how close the system already is to changing state.

So next time you need to measure something almost too small to see, don't first ask "how do I build a bigger amplifier?" Ask "what system is already balanced on a knife-edge, where my tiny signal would tip it into something loud?"

That is how you weigh a nuclear stockpile from across the room: by finding a thermometer sitting so close to the edge of superconducting that the heat of a single photon knocks it over—and reading, from how far it fell, exactly which photon it was.

Distilled from Phys.org