James ActonAnalyzing NORK U

Jeffrey set me a challenge by last week. Before outlining it and my answer, some background might be in order first…

As he outlined in a recent post, in their November draft declaration, the North Koreans continued to deny the existence of a uranium enrichment program. But, in response to US queries, they were willing to provide the US with access to some equipment, including aluminum tubes.

In late December, Glenn Kessler reported in the Washington Post that samples of uranium had been found on these tubes. In the same article David Albright was quoted as saying that this wasn’t proof positive of an enrichment programme. For instance, the uranium could be the result of cross contamination from material supplied by Pakistan. Then, in the Nelson Report of 18 January (relayed by Jeffrey) it was reported that

[T]here is emerging scientific consensus (again) that whatever else the DPRK is actively doing to produce nuclear weapons, it is plutonium based, and that resolution of the HEU question, while important, is not central to the main mission of the 6 Party talks.

Jeffrey’s challenge to me was this: To take apart the science of what they might have done and found, and explain the possibilities and limits of this science in a policy context. Having been marking essays I set all month, it seemed only fair to do his.

Perhaps the most significant aspect of this whole story for uncovering what went on is the timing. Kessler implies that the US knew about the uranium contamination well before his article was published on December 21. So, in all, it could have taken about 6 or 8 weeks from the discovery of the contamination to the conclusion that it was not of central importance. This timescale is interesting because it is about the time it takes to get results from a nuclear forensics technique called fission track thermal ionization mass spectrometry, more usually known as FT-TIMS.

When swipe samples are first brought in for analysis they can be checked quickly but crudely for the presence of uranium using a variety of techniques. However, these techniques don’t reveal the properties of individual uranium particles. Attribution usually requires knowing the isotopic composition of individual particles, including the abundance of the trace isotope U-236.

Only one technique, FT-TIMS, can provide this level of detail. FT-TIMS is a two step process. First, samples are embedded in a plastic and irradiated with slow neutrons to induce fission of U-235 (but not U-238). Fission products leave tracks in the plastic enabling particles rich in U-235 to be identified. Finally, particles of interest are isolated and their isotopic composition identified using standard TIMS. This whole process normally takes several weeks at least.

My guess is that what happened was this: In November or early December the US received the tubes and bulk measurements revealed the presence of uranium. This set off alarm bells. Six or eight weeks later the FT-TIMS results came in and calmed nerves.

So what did those FT-TIMS results reveal? Again, it’s only a guess but I suspect they validated the suggestion (reported by Kessler) that the uranium originated from Pakistan. The IAEA was provided with samples of Pakistani centrifuge components when it was trying to untangle the history of Iran’s centrifuge programme (para 12 of GOV/2005/67). It’s entirely possibly (indeed, very likely) that the US obtained the results of the analysis on Pakistani centrifuges. My guess is that the US has now discovered that the uranium found on North Korean tubes matches that found on Pakistani centrifuges. This would suggest—but not prove—that North Korea has never got as far as running a centrifuge successfully. It doesn’t prove it because we don’t know whether the North Koreans have been selective in what they have shown to the US.

So, to address the second part of Jeffrey’s question, how useful is FT-TIMS in a policy context? Well, if FT-TIMS tells you that something is there then you can believe it. The converse is not true: absence of evidence is not necessarily evidence of absence. That is, if you are looking for uranium particles with particular isotopics and don’t find them, it might well be because you have missed them not because they are not there. IAEA investigations in 2004 into Iran’s centrifuge program demonstrate this nicely. Throughout that year, the IAEA was able to report the discovery of uranium particles with new isotopics. It didn’t get them all in June when FT-TIMS results were first reported.

This is unlikely to be a problem with North Korea. If (and it’s a big if) the US is satisfied that the contamination from North Korean centrifuges came from Pakistan, then it has likely discovered the presence, not the absence, of particles with distinctive isotopics. Further investigations are unlikely to change this conclusion.

So, Jeffrey, does that answer your question?


  1. Lao Tao Ren (History)

    The larger question that this issue raises is the established links between Pakistan and DPRK whom the NKs have now graciously confirmed.

    What else was exchanged beside some shipments of aluminum tubes? What about the Chinese knowhow that went into the Pakistani program.

    How much Chinese (broadly speaking, including Taipei) knowhow ended up in this bazaar?

  2. Eli (History)

    Would North Korea have the ability to produce its own feedstock to put into a small amount of test centrifuges? Or would it have to purchase this from Pakistan along with the other equipment? Just as Iran originally had to purchase their feedstock from China.

    Thanks for this postulating. There really seeemd to be a general gap in understanding as to why uranium contamination in the tube sample didn’t raise more red flags.

  3. Jeffrey Lewis (History)

    I think that is a pretty decent answer, yes.

    I suppose this is why I let you blog on the site.

  4. Hairs (History)

    I don’t understand why it is necessary to allow 6 – 8 weeks to pass for the development of fission tracks. The inherent accuracy of isotopic ratios using TIMS is about 1.2 × 10E-10 (please refer S. RICHTER, A. ALONSO, W. DE BOLLE, R. WELLUM, and P.D.P. TAYLOR,
    ‘Isotopic “fingerprints” for Natural Uranium Ore Samples,’ Int J Mass Spectrom,
    193, 19 (1999)). Hence I see no reason to delay the isotopic analysis, unless the constraint is financial or lack of analytical resources. Instead, just take what particles you have and measure their spectra immediately!

    A further doubt I have is related to the need (wisdom?) of performing a fission track study. Inducing fissions in the U-235 by bombarding it with thermal neutrons is going to generate prompt neutrons along with the fission products. These prompt neutrons DO fission U-238, and additionally may have appropriate energies (or thermalise sufficiently) to be captured by natural traces of U-232, U-233, and U-234 by (n,gamma) reaction. In the case of U-234 (n,gamma) you’re going to end up with U-235, which is exactly what you were trying to measure in the first place. Thus the very act of irradiating the samples would seem to increase the uncertainty in the final analytical results.

    Another reason for NOT irradiating whatever you do have is the effect on the U-233 and U-234: U-234 has a shorter half-life than U-238, and therefore exists in nature more or less in secular equilibrium with U-238, with their relative abundances pretty much the same as their relative half-lives. Thus natural uranium has about 58ppm U-234. HOWEVER, local geology affects the ratios slightly, and consequently there are natural variations of about +/- 3ppm in the U-234 / U-238 ratio. Therefore our esteemed nuclear detectives are probably going to be looking at the ratio of U-234 to U-238 (after accounting for the degeree to which the U-234 would have been enriched) in order to “pinpoint” where the original uranium ore was mined. Thus to mess with the U-238 and U-234 signatures by thermal-neutron irradiation (which leads to fast-neutron irradiation from the fissioning U-235) really makes no sense to me.

    Of course, the relative abundances of other trace metals and fluorides in the tubes will help to constrain the degree of similarity between the particles found in North Korea and those found previously in Pakistan. But even if the signatures are identical it doesn’t tell you (as far as I can see) whether you have traces of an enrichment that was produced in Pakistan, or whether you have an enrichment that was produced in North Korea, but which used the same feedstock as was used in Pakistan.

    In short, even if the North Korean stuff is identical to what was scraped off of Pakistan’s centrifuges, it doesn’t prove that it came from Pakistan. Instead it only demonstrates that it’s the same feedstock enriched in the same way. Thus an isotopic analysis is unlikely to be categoric about WHERE the enrichment was done – and the question of where it was done seems to me to be the heart of the issue.

    …and I’m still baffled by why anyone would want to bugger up the isotopic ratios by inducing 2 months’ worth of fissions first. May I suggest a write-in competition: The first correct answer out of the hat will win a free trip to North Korea to swim in the newly-opened-to-the world Magnox Rod Thermal Spa, with underwater mood-lighting provided courtesy of Dr. Cerenkov. 🙂