Jeffrey LewisDropping the F-Bomb on the NORKS

Well, the French are the first to drop the F-bomb on North Korea—failure.

Speaking to reporters, French Defense Minister Michele Alliot-Marie (right) said “it was an explosion with a force of about half a kiloton, which is not an extremely powerful explosion, or it shows that there could have been a failure.”

AP’s Katherine Schrader paraphrased the official statement from the US Director of National Intelligence simply noting the subkiloton yield was ahistorical. Would it kill AP to run the full quote?

The Intelligence Community detected a Sub-Kiloton explosive event in North Korea. We cannot confirm if it was a nuclear explosion. For an initial test a yield of several Kilotons has been historically observed. The NK Mission Manager will continue to monitor and gather analysis throughout the day.

See, full quote run; no harm done.

Anyway, IRIS has the seismograms online. And no, that is way beyond my skillset.

AFP has a very competent summary of the debate about the test. One possibility is that North Korea faked a test, with a huge amount of conventional explosives—about 1,000 tons, come to think of it. That seems unlikely to me—after all, why fake an explosion that casts doubt on your nuclear capability? Did I mention that was a lot of explosive?

Anway, we should know from radionuclide monitoring in a couple of days.

Update: Kevin Drum brings in a big gun with the chops to deal with data that makes me squeamish.

More Updates: I don’t mean to countenance the idea that the test was staged—after all, why stage a failure?


  1. Jeha (History)

    You never know with those guys…

    you’re right, the next few days will clear it, be it seismic data or nucleide counts.

    If it is fake, the audience cannot possibly be external, but internal. A sign of weakness on both counts.

    Still, a Kilotonne or so is not too bad, with Seoul that close…

  2. Brent Logan (History)

    “Why stage a failure?”

    How about if NK claims to have developed suitcase nuke technology? That’s certainly scarier than the commonly-held belief that NK’s nukes are too big to fit on their missiles.

  3. Bill Arnold (History)

    Staging a failure requires less conventional explosives. Much easier to hide, cheaper, and nitrates are best used on crops in a country that is always at best on the brink of starvation.

  4. Grumpy Physicist (History)

    One point that I think might be relevant: a few days before the Trinity test (1945, the first nuclear explosion), there was set off a test of the implosion device, with a non-fissile core.

    It was just to see if the implosion worked, since the implosion was the difficult part of the Fat Man design. This was known as the “Creutz charge” (look it up in Rhodes, ‘Making of the Atomic Bomb’), and was about 5kT of high explosive.

    Something like an implosion test would be consistent with both the available seismological evidence, and the NK statement (“a successful nuclear test, with no radiation released”)

  5. nick s (History)

    “I don’t mean to countenance the idea that the test was staged—after all, why stage a failure?”

    Because the American media (and more importantly, the White Hose) won’t report on it with the same immediacy and detail? It’s not in certain political interests to cast NK as having either a dud test or a staged test. The intention to test is all that matters.

  6. Steve

    That doesn’t make any sense. Implosion tests of the firing set without fis-mat don’t produce anywhere near this big a seismic signature. What kind of nuclear device is designed to use a minimum of 550 tons of conventional HE to initiate the chain reaction?

    Besides, the DPRK has been doing conventional explosive tests, presumably of firing sets under development, since the early 1990s.

  7. Steve

    The Trinity design did NOT contain 5,000 tons of conventional HE. I don’t think the B-29 that delivered it even weighed that much. I think the entire “Fat Man” bomb only weighed about 7-10 tons, don’t have references handy. And if the yield of the bomb dropped on Nagasaki was about 10 kilotons, that would mean fully half of it came from conventional explosives, if this figure is correct. Something must be wrong in this assessment.

  8. AZ_Squeegee (History)

    This paper in the Bulletin of the Seismological Society of America seems to indicate (on page S151) using data obtained from previous nuclear tests of known yields that the estimated magnitude of 4.2 for the shock from the explosion indicvates a much larger yield than 550 tons of TNT.

  9. Richard

    Is it possible DPRK had to use reactor-grade plutonium, and accept the higher risk of a fizzle (predetonation)?

    When the UK used similar Magnox reactors to produce weapons-grade plutonium, they changed the fuel load twice a year. The 2005 Siegfried S. Hecker report suggests the latest Yongbyon 5MWe fuel load was in for two years (Feb 2003 – Mar 2005) on a full power run, which seems like it wouldn’t produce weapons-grade plutonium.

    It has always intrigued me that the publicly released Yongbyon reactor info doesn’t discuss the grade of plutonium produced, which of course careful satellite monitoring could estimate.

  10. blob


    One question: is it possible to dig, in secret, a big cavity which could reduce the shockewaves of a 10kt nuclear explosion?

    I have read that it was proposed to hide nuclear tests.

  11. Grumpy Physicist (History)

    My mistake, it was 5000 lbs (one gets so used to everything in tons) not 5000 tons.

    So about 2.5t for the Creutz test. Yes, much too small to explain the seismic event in NK, and I have to agree with the question in the original post: why fake a failure? And so far, it looks an awful lot like a failure.

  12. Andrew Foland (History)

    My copy of Rhodes’ book says the Creutz shot was 5,000 pounds, not 5,000 tons. This seems to be more plausible.

  13. Thomas Mattison (History)

    Before pronouncing the test a fizzle, at least two things should be remembered.

    First, we don’t know the seismic calibration offset for the NK test site. I seem to recall that there was a factor of 2 or 3 difference in seismic energy transmission for Soviet test sites vs Nevada test site.

    Second, if the test is done in a large (tens of meters radius) cavity, most of the energy can be decoupled from the ground. The energy just heats and thus raises the pressure of the air in the cavity. This doesn’t give much of a seismic signal. The US got a factor of 70 decoupling for a sub-kiloton device in a large cavity (left by a previous test). The Soviets made an 8-10 kT device look like a sub-kiloton device (magnitude 4.1) by testing inside a 37 meter radius cavity.

    See National Academies of Science Press

    Technical Issues Related to the Comprehensive Nuclear Test Ban Treaty (2002), page 46

  14. Alex (History)

    “was about 5kT of high explosive.”—This isn’t correct. The high explosive in the test implosion bomb was simply a standard set of explosive lenses, way less than a kiloton in strength (though the “Pumpkin” bombs, even without the plutonium core, was still a pretty powerful conventional bomb). Perhaps you are confusing this with the 100 ton test of TNT that took place before the Trinity test to help calibrate the instruments.

    There are some simulated nuclear explosions (i.e. the “Minor Scale” test of 1985 that the US undertook) which are in the low-kiloton range (4 kt in that case) but those were for a pretty specific purpose.

    In the end, it seems highly unlikely to me that North Korea would bother to fake a fizzle. Perhaps the Illest is trying to use Juche mind tricks on our feeble Western brains.

  15. Dave (History)

    I think this test has everyone crapping in their pants and talk of a fizzle is pure disinformation. It is about the size of the first stage of a thermonuclear device and would use just a few ounces of plutonium. The real question is whether it was a multi-stage device in which case the amount of plutonium used could be smaller. Modern 500 T devices can be made to fit into something the size of a normal dictionary and are virtually undetectable.

    The nokos are smart guys and have had 50 years of examples to work from. I would expect them to produce a modern device on the first try.

    No I am not a nuclear weapons designer, just an ordinary engineer with 40 years of design experience.

  16. j house (History)

    Is the bomb physics harder to produce a sub-kiloton (and critical) explosion?Politically, the ‘look at me’ strategy would necessitate producing a larger yield explosion, so the notion of a dud isn’t far-fetched. We’ll see.Perhaps they were testing a smaller, advanced warhead that is missile capable, and failed.One would think they would already be confident that their old, reliable Chinese bomb designs needn’t be tested.

  17. Tom (History)

    One quick point/question to bring up with regards to weapons design by the North Koreans…

    Wasn’t the DPRK the alleged supplier of the plutonium used in Pakistan’s 1998 test of a boosted fission weapon? After all, that test was roughly three years before Pakistan is assumed to have been capable of separating plutonium indigenously.

    Just to play devils advocate for a moment if North Korea was cooperating with Pakistan through the mid-nineties on weapons design isn’t it slightly possible their program is more advanced than a purely DPRK effort would be?

    Perhaps they were attempting to test the low end of their yield range for weapons design.

    That’s would certainly fit in with their attitude towards propaganda and disinformation. Don’t test something you know works as that reveals your capabilities.

  18. Yale

    Even if the test was a “fizzle” it was NOT a “dud”.

    It simply predetonated a microsecond early. Another test or two (with or without fissiles) will correct it.

    A 1/2 kiloton “fizzle” is not a toy. It is an horribly destructive weapon.

    It showers a lethal dose ( >500 REM ) of radiation over an area of 2 square kilometers!

    The major cities in the region have densities of 10K to 30K persons per square kilometer. At times (business zones at lunchhour, festivals, etc) the density is 10-20 times as high.

    The lethal level blast and thermal pulse would also blanket a 1 square kilometer area. The fallout would contaminate kilometers downwind.

    The US had thousands of nuclear weapons capable of yields less than one kiloton.

    The tactical version of the B-61 bomb has a yield which includes 1/3 kiloton.

    Even if never optimized, the NKs produced a horror from hell.

    Here is an image of a US bomb test of a device with 1/3 the yield of the NK test:

  19. Thomas Mattison (History)

    From the:New York Times

    “A senior Bush administration official said he had learned through Asian contacts that the North Koreans had expected the detonation to have a force of about four kilotons. ”

    If we take the 4 kT expected yield seriously, that means that rather than a conservative design like a uranium gun-bomb, or solid—core implosion bomb, both of which have several normall-density critical masses of fissile material and give you 10-30 kT, the NK design was an advanced design like a levitated core. These use less fissile material, so they have a lower yield. The yield is plenty for igniting the secondary of a hydrogen bomb, which most US and Soviet bombs are. So you can make more bombs per kg of fissile material. Even as a straight fission bomb, a small advanced design is desirable if your only ICBM has a low throw weight.

    There’s enough info floating around between the internet and AQ Kahn that a you can have reasonable confidence in a conservative high-yield design without testing. But if you are trying to be tricky to stretch your limited fissile material, you would want to test.

    I don’t believe anyone who says that the margin of error on the yield estimate for a test at a site with unknown local geology is better than a factor of a few.

    So an estimated yield of about 1 kT on a 4 kT design is just as consistent with complete success as it is with a fizzle.

  20. Rocketman

    The reason the French are playing this one down is because Chirac is in bed with the Iranians and Norks.

    ‘DUD’ !! !! ?? ??? ????I love how those GOP fans try to put a spin on everything, the DPRK dictator has always openly said he’s got a WMD program ( bioweapons, nuclear and chemcial )It ain’t a dud, anyone who says it to be a ‘dud’ clearly have no understand of nuclear tesing. Many of the British nukes at Operation Antler, the US Nukes at the Nevada test site, and the Pakistani Nukes produced very small yields (but would still be powerful enough to kill hundreds of thousands people).

    Perhaps when that dictator Kim finally gets to build his ICBM with all those cherry-on-top warheads he desires so much, so he can turn SanFranciso and Texas into radioactive craters we will finally be convinced this madman has got a few Nukes !

  21. jwilliams

    On May 7, 1945 the united states tested a 108 tons of TNT in preperation for the Trinity test.

  22. sunbin (History)

    great post and great discussions.

    “why stage a failure?”

    one plausible explanation is that the scientists were not ready, but KJI pushed for it for its own peculiar political agenda.

    why failed?IMO it should either be the Pu grade/purity or the implosion device. I have been skeptical of NK technology since the TPD fiasco in july.

    Note that China had only been closed for 15 years when it produced the bomb, with scientists either involved or close to Manhattan project heading the team. NK does not have such expertise nor such a talent pool (600M population base vs 24M).—-p.s. SK intelligence had failed both in july and again this time. they seem to be from a different university than the Samsung team.

  23. Arrigo (History)

    OK, I’ll go out on a limb here: there were reports (from the SK intelligence?) about the tunnel being 700m long as opposed to the expected 400m or so “expected” for a “standard” 10kT blast, all based on the amount of earth piled up near the tunnel.

    Well, what if the mounds of earth were not 700m of tunnel but 400m plus a large decoupling cavern?

    Then the formula which gives you the 500T fizzle suddenly no longer holds and it becomes quite feasible for it to be something larger.

    The other question I have is: where do people get the geology info for those mountains? Has the DPRK recently revealed what’s 400/700m into those mountains? That is also very relevant w.r.t. using the correct formulae.

  24. Rocketman (History)

    A few GOP diehards will claim it was merely a big bundle of dynamite going boom and looks like you can always count on the French to act like idiots. Maybe the GOP will finally believe Kim has Nukes when he sticks a warhead on top of that ICBM he’s trying to make and turns LA into a radioactive crater.

  25. Alex (History)

    “The nokos are smart guys and have had 50 years of examples to work from. I would expect them to produce a modern device on the first try.”

    They’ve managed to bungle their missile program up pretty good, and whatever their experience with weapons design, I suspect their experience with plutonium (“An engineer’s nightmare, a physicist’s dream” as one Los Alamos publication puts it) is probably not to the level of what the US managed to have in the 1960s when it was able to make its first very small bombs.

    “Is the bomb physics harder to produce a sub-kiloton (and critical) explosion?” In the case of the experience with the US and Soviet programs — which are not necessarily the benchmark here, since those were in the 1950s and 1960s and here we are in the 2000s — the design of highly miniaturized weapons, the sort which detonate a few kilotons in yield and can be just-barely man-portable (like the W54), the answer seems to have been yes. I don’t know the details but I imagine it has to do with the fact that a critical mass of nuclear material is created it generally wants to detonate somewhat larger than a few kilotons (and while you could purposely fizzle that would be pretty messy and wasteful of valuable material), and that the geometries of small weapons are more difficult than those of large weapons (and I imagine you have less room for error). But I’m no engineer or physicist or anything like that, but the historical record (which I feel more competent about) seems to indicate that making fission weapons small (both in physical size and in yield) was a major difficulty and a major area of research at the US labs through the 1950s and 1960s.

    Here are the possibilities suggested:

    1. North Korea has set off a dud for some reason (bad plutonium seems like as good a guess to me as any).2. North Korea has really set off a highly sophisticated advanced warhead.3. North Korea has set off a sophisticated warhead which is a primary for an even more sophisticated weapon.

    Now given that it is North Korea we are talking about here, a place which has a pretty lousy track record for testing things successfully (i.e. the missiles) and a pretty good track record for amazing failure (i.e. their industrial accidents they seem to have on a regular basis, not to mention that whole starvation-country-who-spends-most-of-its-time-making-fake-US-currency thing), I’m inclined to think that option #1 is the one which requires the least leap of fancy at the moment, all other things considered. I also find it unlikely that North Korea wouldn’t want to set off at least a “nominal” fission weapon for their first test — tradition makes it seem like a good idea, for one thing, but since they obviously hope this test will be their ticket to a better deal of some sort one would expect them not to want people to be speculating about whether or not the rocket in their pocket is a little limp.

    Now I don’t claim to be an expert on nuclear proliferation or North Korea in the slightest but I find the dud story more compelling than the others, which require North Korea to be both confident in its international position as well as its technical position. I’ve not really seen evidence of either of these things. None of this is meant to be some sort of Western chauvinism or to imply that we shouldn’t take this seriously or not.

  26. Alex (History)

    Also, even a very successful faked nuke would look only like a very low-yield nuclear bomb (i.e. low kilotons at best), getting us into the whole “was it a dud?” question no matter what. I find the possibility of them failing a fake humorous but extremely unlikely, mostly because I find the idea of them faking it very unlikely to begin with. Out of all of the options, “they tried to set off a nominal bomb but failed” seems most likely to me, but that’s mostly just my intuition speaking.

  27. Antonio Manetti (History)

    Is there any point in comparing or correlating the seismic data with the signature from known nuclear underground tests?


  28. Yale

    It is completely incorrect to call the test a “dud”. A dud is detonation where the nuclear yield does not exceed the explosive force of the high explosive shell surrounding the core.

    A “fizzle” on the other hand, is a nuclear yield lower than designed.

    A 1/2 kiloton nuclear “fizzle” blast is a fearsome nightmare, far, far, far exceeding any other conventional weapon. It would destroy essentially EVERYONE over a two square kilometer area with a radiation pulse and shatter with blast and heat a one square kilometer area.

    We must not trivialize this.

    The next test by the NKs will quite likely go nominal yield (maybe 4 kilotons), but even a fizzle should not be laughed at.

    Don’t forget, the NKs were capable of creating an implosion precise to less than a microsecond. They are not clowns.

    Tweaking the design to maintaining an assembly for only an few dozen NANOseconds more will be sufficient to go full yield.

    There has been discussion about NK using minimum plutonium in order to minimize the weapon size. Precisely the opposite would be true in an early generation missile-capable warhead. A kilogram or so of Pu either way is trivial. It is the weight of the conventional high explosives (HE) and other components that create the volume and mass problem. The more Pu in the core, the less HE and the smaller the warhead. If Pu is in short supply, you go the other way, minimal fissile and heavy HE.

    Advanced generation designs use both minimal HE and minimal Pu, but that would not be relevant to NK.

  29. Lab Lemming (History)

    The Indian and Pakistan tests were around mag 5. A Chinese thermonuclear test last decade was 6.6, and the largest underground US test was the 5,000 kiloton Cannikin test, with a bodaywave magnitude of 6.9

  30. linuxguru (History)

    1. If their Pu is contaminated with heavier isotopes as well as light elements (like Al, Mg, etc.), then predetonation is unavoidable, regardless of the precision of the HE assembly and timing of the initiator – a fizzle would be the expected outcome, though the magnitude would vary stochastically depending on when a stray alpha hit a light element nucleus, while approaching criticality.

    2. They have moved around multi-ton quantity of ammonium nitrate in the past, including on a freight train that blew up a few years ago. It’s conceivable that they could fake a sub-kiloton device with conventional explosives, if given enough time to hoard it in a cavern.

  31. Yale

    inuxguru wrote:“If their Pu is contaminated with heavier isotopes as well as light elements (like Al, Mg, etc.), then predetonation is unavoidable,”

    That is not correct. ALL Pu used in weapons contain unwanted neutron emitters. The likelihood of preinitition for any level of the intensity of these emitters is dependent on the specifics of the weapon design – (implosion velocity being the most important.)

    And with boosting, preinitiation is irrelevant. The boost occurs at less than 1/3 kiloton.

    From the weapons designer Ted Taylor:

    “Utility of Reactor Grade Plutonium in Nuclear WeaponsJune 10, 1998

    Contrary to opinions expressed by many nuclear engineers that are not familiar with the still secret intimate details of nuclear weapon design and operation, plutonium extracted from all types of spent fuel removed from nuclear power plants or research reactors can be used for making modern fission or thermonuclear weapons that are reliabily predictable in performance, over a very wide range of yields, from fractions of a kiloton to megatons of high explosive equivalent. This has been true for decades, and confirmed by numerous nuclear weapon tests.

    It is true that the first generation of implosion type fission bombs, such as the one that destroyed Nagasaki in 1945, could “fizzle” and produce much lower than the design yields if the plutonium they contained were of “reactor grade.” This could be the result of premature initiation of a fission chain reaction by spontaneous fission neutrons emitted by Pu-240 or other isotopes that are more abundant in reactor grade than weapon grade plutonium. But ways to avoid this problem, by use of plutonium in different designs that could be reliably used for fission and thermonuclear weapons were developed and demonstrated before the end of the 1950s. The performance of these weapons is not significantly degraded by using reactor grade plutonium instead of weapon grade plutonium.”

    Discussions of the Korean bomb imply that it is a second generation design (levitated core, etc)