Joshua PollackNorth Korea's Indigenous Bomb Design

It’s a red-letter day for Norkologists. For one thing, Kim Jong-il’s in Beijing, taking what’s almost certainly his last train trip outside the country. According to the LA Times, he’s dispensed with the usual pair of decoy trains. Perhaps he’s now close enough to the end of the ol’ personal rail spur that he’s no longer quite so afraid of death. Or maybe the DPRK is just out of cash.

On a completely different note, the May/June issue of the Bulletin is out, with no fewer than four NK-related articles. (All are behind the paywall.)

Now, don’t worry. I only plan to talk about one of them.

Buddy, Can You Spare a Warhead Design?

In An alternative view to North Korea’s bomb acquisition, Li Bin of Tsinghua U. fame offers a provocative argument: that North Korea’s apparent decision to build and test a nuclear device with 4 kt yield — unusually small for a first weapon — suggests that they were using a bomb design originally developed in some other country:

If we combine two of the main assumptions proposed above—that the North Koreans aimed for a 4-kiloton yield and that they could not complete such a design without the benefit of previous experience—a potential workable explanation emerges: Pyongyang had access to an existing nuclear weapon design.

(Readers may notice some similarity to Geoff Forden’s recent essay on the North Korean missile program.)

Choose Your Anonymous Report

While we can’t rule out the possibility of foreign assistance, it doesn’t appear to me that there’s compelling evidence for an existing design. (Before you ask, the Chinese-origin blueprints delivered by the Khan network to Libya — in a bag from a Rawalpindi dry cleaning shop! — were for an HEU-based weapon. North Korea is working with plutonium.)

Li’s assumption that the device was too sophisticated for beginners is based on a report in the South Korean media that the DPRK’s 2008 declaration claimed that only 2 kg of plutonium had been used in the October 2006 nuclear test. But it ain’t necessarily so. A leaked defector report in the South Korean media in 2005 described an initial device with twice that much fissile material, and some questions about its performance:

Kim [a pseudonym] reportedly said that “North Korea manufactured a nuclear weapon weighing one metric ton with four kilograms of plutonium” and testified that “although North Korean scientists have reported to Kim Jong Il that the nuclear weapon will performance [sic] normally, they in fact have doubts about the performance of the manufactured nuclear weapon.”

Since the 2006 test appears to have fizzled, the doubts attributed to North Korean weapons scientists seem to have been well-merited. The fizzle itself is reason to suspect that it wasn’t a proven design.

But again, a great deal depends on whether you prefer the 2 kg Pu claim from 2008 or the 4 kg Pu claim from 2005. It just so happens that the yield curves from the Cochrane and Paine paper show that a “low tech” bomb designer ought to be able to get a 4 kt yield from 4 kg Pu, and a “medium tech” designer ought to be able to get a 4 kt yield from 2 kg Pu.

Until more evidence emerges, my money’s on a juche nuclear device.

For background to this post, see what Jeff was blogging back in April 2005, July 2005, and October 2006. Yes, this blog is that old now…


  1. Azr@el (History)

    After reading Forden’s recent paper on the DPRK’s missile program…I can only say the North Korean Rocketeers must be bloody geniuses of the highest caliber. If you asked the greatest rocket scientists of Russia and US to get a missile like the SS-N-6, a missile to designed to be launched flooded with water to a depth of nearly 50 meters to launch in near vacuum as a second stage as Forden suggest occurs in the 2nd stage of UNHA 2, I think they would all laugh and tell you it was impossible. But those Plucky North Koreans managed to pull off…Kudos to them!…or they just simply used a scud or nodong engine with a simple extender… whichever occam’s razor favours.

    With respect to the DPRK nuclear program, I think it’s highly probable that no one will try to reinvent trinity, fatman, etc… The physics of condensed matter and neutron transport is too well known, the engineering skills too readily available, every would be proliferator will shoot for a medium to advanced design and scale back to their means. The bottleneck is fissile material and trit not design.

  2. Allen Thomson (History)

    Let me take the opportunity to solicit opinions about where the design would have come from for the bomb that would presumably have been made with the plutonium from the now-former Syrian reactor.

  3. John Field (History)

    FWIW, I lean to the 2 kg medium tech explanation.

    The strategic advantages for the NORKs having a limited supply of plutonium would be very large. You get twice as many bombs and they would weigh half as much. So, you might be able to get them on a missile. Meanwhile, damage would scale as the 2/3 power or something. Meanwhile, if you had uncertainty about yield, you could deliberately target lower than optimum detonation altitudes to mitigate the uncertainty. And, if you were worried about a fizzle, of course, you now have half as much plutonium around too.

    The bare critical mass is 9-10 kg. With a reflector, around 5 kg or so. Increase the density by 2x and the critical mass drops to about 1.25 kg. LASL Shock Hugoniot manual suggests to me that this is very doable.

    This all looks to me like an argument that the NORKs were telling the truth at 2 kg. But, its going to require that they get the Pu moving pretty fast which implies that the pit needs to be fairly thin. And that means it is hard – not impossible.

    I go with a juche device too.

  4. PC (History)

    Stillman and Reed also suggested that the Norks used the CHIC-4 design, adding that one explanation for the 06 fizzle was the inability to properly convert an HEU design to use Pu. Who knows…but for a technical layperson like myself, that seems plausible enough to be a possible explanation for the motivation to go to 4kt and the outcome. Though I’d be curious, if that were technically plausible, if it would make the 2kg Pu device less likely since you’d already have to work with a different mass of material and scaling that mass down so much further would be far more of a technical challenge than its worth.

    And foreign assistance wouldn’t be inconsistent with a juche bomb in practice—since the real world application of juche apparently means you’re totally self-reliant…on international aid to eat.

  5. Josh (History)


    I didn’t make much of it, but since you raise the issue of halving the weight, here’s the immediately following bit of the defector report:

    Kim reportedly also testified that “since North Korea does not feel confident whether or not large nuclear weapons will explode to its capacity in actual warfare, it has recently been in the process of manufacturing a nuclear weapon miniaturized to 500 kilograms.”

    Make of that what you will.

  6. anon


    I’m not sure 2 kg vs 4 kg of WGPu translates to halving the weight of the device. There may be some savings because there is not as much Pu to compress, but now you need to compress the 2kg more. Maybe I’m wrong, but that connection is not obvious to me.

  7. John Schilling (History)

    The Cochrane and Paine paper is missing one key variable – the size (mass and diameter) of the implosion + tamper assembly. It is most definitely not the case that by using half the plutonium “you get twice as many bombs and they weigh half as much”. More likely, you get twice as many bombs and they weigh twice as much.

    Or maybe you get twice as many bombs with the same weight and one-quarter the yield, or twice as many bombs at half the weight and one-tenth the yield, depending on the weapon designer’s priorities. But cutting down on plutonium means you really want to wrap what plutonium you’ve got with a thick, heavy tamper, and enough HE to squeeze it extra-tight. Note that the US Mark 12 atomic bomb reportedly used significantly more fissile material than the heavier, higher-yield Mark 7.

    The North Koreans certainly want to minimize their use of scarce fissile materials. They also want warheads that will actually fit on their missiles. And they want warheads that make explosions large enough that people won’t laugh and point and say “that’s not a nuclear weapon; this is a nuclear weapon!”. Would be nice if we had a chart that actually covered the whole trade space.

    My best guess, is a ~4 kiloton warhead weighing ~500 kg and of ~70 cm diameter, using ~4 kg of plutonium. Designed by moderately capable “medium technology” weapons designers; I don’t know and don’t much care whether they are North Koreans, Pakistanis, or Chinese.

  8. George William Herbert (History)

    The chart John Schilling wants is multidimensional and hard to put on paper. And as much as I encourage open discussion, would make me nervous to put out there.

    Even ignoring lens systems, just the implosion assembly performance information for the supercharge and pit, with varying pits, is really sensitive.

    It’s sort of annoying, because one could (cough) theoretically generate it with a number of different modeling assumption scenarios with an Excel spreadsheet, for various explosives, supercharge outer diameters, pit diameters, and pit configurations and weights, with essentially any variety of variations one wanted. It’s not like spherical implosion codes or the spherical implosion Gurney Equation are that hard, and the EOS for the various materials aren’t that hard either (yes, Plutonium’s is partly classified, but even if it hadn’t leaked around the edges the Fermi Gas approximation is just too easy…).

  9. Allen Thomson (History)

    The other thing that’s worth considering is the possibility that the DPRK would be willing to make bombs that would be too big/heavy/inefficient to fit on top of a missile or for military use in general — but would be deliverable by truck/boat/airliner etc.

    And, as the NRDC paper notes, even a 1 kT bang is a pretty big bang.

  10. Azr@el (History)

    I recall reading a Russian paper on hyper implosion of plutonium or uranium targets. The theoreticians felt they could get 300 times compression, i.e. critical mass of 180mg using shockless compression via heavy ion beams.

    The energy requirements to compress a certain mass of fissile material seems to shoot up near a power of 2 as a function of compression, thus I’d assume the mass of an isentropic implosion assembly for x compression to be (I*K*x)^c; I is efficiency of the driver, K some constant to account for unit system chosen and c is a function of the critical mass for a given value of x..i.e. a function of x that tapers down from 2 as x—> oo. The energy to compress a mass of fissile material for a given level of x increases faster than the reduction in energy from compressing a smaller fissile mass for a given x. Thus a small implosion (1.2-1.3 times compression) of say 10kg of pu239 may only require an assembly of 10 kg, whereas an implosion on the order of 2 times compression will require <100kg and a facility capable of compressing pu239 to 300 times will require a few hundred tons.

Pin It on Pinterest