What to do when you want to perform physical tests of the plutonium in your nuclear weapons and you've pretty much told the world you won't set off any more nukes in these kinds of experiments?

If you're the US government, you can do what's called sub-critical tests, the results of which can help scientists understand what will happen in a fully fledged detonation.

There is a ton of literature out there on how nuclear weapons work, but as a brief overview: a primary component of today's weapons is typically the radioactive metal plutonium that, at the moment when the device needs to go off, is rapidly compressed to such a density that it goes from sub-critical to critical, allowing a fission chain reaction to be sustained. The compression is achieved by setting off carefully shaped conventional high explosives placed around the plutonium, which crushes the metal core.

This sustained, uncontrolled reaction by itself can be rather destructive: the atomic bomb that was dropped on the Japanese city of Nagasaki in World War II used this plutonium implosion and subsequent chain reaction to produce a blast wave, heat, and radiation that devastated a good proportion of the city. It's estimated that Fat Man bomb exploded with the equivalent force of about 21,000 tons of TNT. In more powerful and complicated thermonuclear weapons, the effects from plutonium compression in a primary stage can drive a secondary fusion stage, in which hydrogen isotopes fuse and release neutrons and energy, which all-in-all results in city-annihilating explosions with the equivalent force of hundreds or thousands of kilotons of TNT.

Fast forward to today, and Uncle Sam's nuclear weapons teams have two things on their minds: how do you check that the decades-old plutonium in America's nuke stockpile still compresses and reacts as expected; and how do you improve the internal structure and design of the nuclear packages within the weapons and know they will work, without setting one off?

You can perform simulations in software, though you need data from empirical studies to confirm the code is correct. For one thing, there are hydrodynamics involved as the plutonium is turned to liquid during its implosion, and modeling not just that but the nuclear physics occurring in that moment is non-trivial.

It's been known for some time that the United States has conducted, and still is conducting, some sub-critical tests of its nuclear weapon fuel. This is usually performed by compressing some non-critical fissile metal in such a way that it remains sub-critical: it doesn't achieve a sustained, uncontrolled nuclear chain reaction. At least some of the plutonium atoms undergo fission and split into byproducts, releasing neutrons and radiation, all of which can be observed, though a meaningful chain reaction and explosion does not kick off, according to the government. One way to do this is simply to not use that much plutonium, we believe.

Uncle Sam argues these kinds of experiments are not a violation of its promise to not conduct overground and underground nuclear weapons tests.

This month US Dept of Energy talked publicly about an $1.8 billion machine dubbed Scorpius it is building to capture images of the compressed plutonium in nanosecond-scale increments using X-rays. That frame rate is needed due to the speed at which the compression and subsequent reactions happen. It is, in a way, a rather expensive photoshoot. The data from these sub-critical experiments will, we're told, be used to check materials simulations are correct.

One of the dept's most important responsibilities is ensuring that should the dark day come that America's nuclear stockpile is needed, the bombs can still do the job, hence its need for Scorpius.

"If you had a car in a garage for 30 to 50 years, and one day you insert the ignition key, how confident are you that it will start," Sandia National Lab's Jon Custer explained. "That's how old our nuclear deterrent is."

The agency largely relies on massive supercomputers to run its nuclear simulations in software, providing insight into how weapon components will perform in the fractions of a second after a device is set off. This is what the newly inaugurated Crossroads super we looked at last month is tasked with. But the question remains: how do you know the simulation is accurate, especially when modeling a half-century-old warhead's capabilities, or an updated version of it?

The DoE calls the process of monitoring fission reactions within a plutonium sample right up to the point just before criticality "tickling the dragon's tail." It's called this because its scientists want to get as close to a sustained chain reaction within a weapon component as they can without actually getting it going in an uncontrolled manner.

The football field-sized Scorpius instrument, key to providing the dept's empirical data, is under construction and will eventually be deployed at the government's U1a Complex in Nevada. It produces images using pulsed X-ray beams, rather than using visible light captured on film or by a digital sensor. The machine will, according to the dept, "be able to produce four separate 80-nanosecond pulses of electrons at 1,400 amps per pulse. Those four pulses can be produced anywhere the experimenters want over a three-microsecond window."

It's a capability Custer said "really puts the computer codes to the test." By computer codes, he means: the simulation software on the supers.

These images will be compared against those generated by supercomputers - such as the Crossroads system we mentioned earlier - to gauge the true state of the United States' warheads.

According to the DoE, ongoing modernization efforts on America's nuclear arsenal will take about 15 years to complete, though with Scorpius, they hope to cut that time down to just five.

"We are entering an era where our modernization programs are going to start making significant changes to the nuclear explosive packages, even if the performance characteristics of the weapons don't change," Daniel Sinars, director of Sandia's Pulsed Power Center, explained.

However, it will be a few years yet before the DoE can start tickling the dragon's tail in this way. First, the dept needs to finish building and testing Scorpius, which won't happen until 2025 at the earliest. Installation in Nevada will take another two years, with the first test slated for late 2027. ®

https://www.theregister.com//2023/10/09/doe_nuke_test/