Permissive Action Links – how nuclear weapons are secured
cs.columbia.edu
cs.columbia.edu
For fission weapons, isn't the most critical part the core? One design criterion for a PAL is to prevent scattering of fissile material, so the core remains intact, right? How hard can it be for a rogue state or group to create a new bomb given that the fissile core is the most valuable part of a fission bomb?
For a fusion bomb, I guess the PAL would damage the aerogel ("fogbank") among other things. That would take some advanced knowledge to redesign, but whoever has the weapon would still have the fissile material from the first stage even if they can't get the fusion part of humpty dumpty back together again.
The original fission bombs were designed in the 1940s with slide rules. It seems to me that PALs do a better job of protecting against unauthorized fusion detonations than of protecting against unauthorized re-engineering of the bombs into fission or radiological versions using the fission core.
I would assume that the US has very few of the ancient "enough fissile material to go critical trivially" sorts of bombs. The miniaturization of atomic weapons to fit into, among other things, ICBM warhead nose cones, likely came about by using smaller quantities of fissile material with much higher precision explosives, timings, and pit shapes to push that smaller nuclear pit into supercriticality. Instead of a soccer-ball of dozens of explosive lenses like the Fat Man, you've got maybe two or three lenses tops. Less moving parts, smaller overall bomb, but much much more timing sensitive.
So if tampering or failing the PAL blows out an EEPROM with the detonation timings on it, you've got a chunk of fissile material that's likely not enough by itself to construct a working nuclear bomb without 30-40 years of superpower-levels-of-funded R&D into materials sciences, precision machining of nasty stuff like beryllium, slapper detonators, so on and so forth.
Most of the gigantic TOP500 supercomputer clusters we've built over the years at national laboratories are likely used towards simulating whether the pits, even in their current configurations, are decaying in a reliable way that the weapons still function if needed. Reprocessing a single given pit into a new weapon is quite possibly an even harder problem.
Now, if you can steal a whole cache of nuclear weapons, then it gets a lot more feasible, since you can probably make one "easy" bomb out of a few hard ones, and either way you have nasty dirty-bomb potential, but I think if you had the capacity to build a working nuclear weapon out of a modern warhead, you probably already have them to begin with.
By some accounts, several incorrect passwords will set off the explosives. Asymmetrically, of course. No nuclear explosion, just a conventional one that kills the fool diddling the bomb and disperses plutonium, making his hideout uninhabitable.