Letter to the Editor: The North Korean Test and the Limits of Nuclear Forensics

Jungmin Kang, Frank N. von Hippel, and Hui Zhang

Harold P. Smith (“Nuclear Forensics and the North Korean Test,” Arms Control Today, November 2006) writes:

Any nuclear explosion creates radioactive noble gases, notably xenon and krypton, that do not combine with other elements in the geologic structure. Therefore, they can more easily leak to the surface and into the atmosphere where they can be detected beyond national boundaries. Because at least two different gases escape, it is possible for radio-chemists to determine if the fissile material was plutonium or uranium, which of course is exactly what happened. (Press reports said the material was plutonium.)

Smith is accurate in writing that the press reported that the material was plutonium, but U.S. officials have not done so publicly. In fact, the office of Director of National Intelligence John Negroponte said nothing about the fissile material used when it announced that it had detected “radioactive debris” two days after the test.[1] We have no doubt that the material was plutonium, and we do not claim to be experts on this subject, but we are skeptical about Smith’s statement that measurements of radioactive noble gases alone can determine the fissile material used in the North Korean test, particularly if detected as much as two days after a test.

The dominant fissile isotopes in a plutonium or a highly enriched uranium (HEU) bomb are plutonium-239 and uranium-235, respectively. We therefore limit our discussion to them. When they fission, various products are created, including several radioactive noble gases. Among these, Xenon-131m, Xenon-133, Xenon-133m, and Xenon-135 are often detected from underground tests.[2]

Smith mentions the radioactive krypton isotopes, but we are not aware that they have been detected from underground tests, and it would be particularly hard to do so from the small North Korean test. Most krypton isotopes have very short half-lives; Krypton-85 is the only one produced in fission with a half-life of more than five hours. In fact, it has a very long half-life (11 years), but it has another attribute that makes identification difficult: it is released in large quantities when spent fuel is reprocessed.

Since Krypton-85 has such a long half-life and spent fuel reprocessing has taken place in a number of countries, large quantities of this isotope have accumulated in the atmosphere. It would be difficult to pick out the small amount of Krypton-85 leaking from a small underground test from this large background. We therefore disregard the radioactive kryptons in the remainder of this discussion.

Moreover, since the amount of the xenons that is released by an underground test is very uncertain, any clue to the nature of the fissile material would have to come from looking at isotope ratios. These ratios are different for uranium-235 and plutonium-239 fissions. The table below shows the fission-spectrum yields of the xenon isotopes that have been detected after nuclear tests in Nevada. We show separately the amounts that are produced directly and those produced indirectly through the decay of iodine-131, iodine-133, and iodine-135, which have half-lives of eight days, 0.87 days, and 0.27 days, respectively. As the chart illustrates, most xenons are produced indirectly, and the isotope ratios from indirectly produced xenons do not differ greatly between plutonium and HEU weapons.

Xenon Isotope Yields per Fission

If one were able to analyze the resulting mix within the first few hours [after an explosion], when the directly produced xenons dominate, it would be possible to distinguish between the xenon from plutonium and uranium explosion[s]. If the air samples were taken two days after the test, however, as Negroponte’s office said, such determinations would be far more difficult. That’s because xenon isotopes produced indirectly through iodine decay predominate, and the ratios of these indirectly produced xenon isotopes do not differ greatly between plutonium-239 and uranium-235 fission (See January/February 2007 print edition of Arms Control Today for accompanying information graphic ). Also, the dilution resulting from atmospheric mixing would make it far more difficult to measure these ratios exactly.

An additional complication is that the parent iodine isotopes of the indirectly produced xenon may not be released from the ground with the directly produced xenons in the same proportions as they are produced. Some iodine might, for example, condense in the ground (the melting and boiling points of iodine are 114[oC] and 184 oC, respectively) or be captured in ground water before the gas from the explosion reaches the atmosphere. This would change the xenon isotope ratios downwind.[3]

Our point here is only to question what can be learned from xenon isotopic ratios alone. We do not question that, if fissile material from the North Korean test was released into the atmosphere by a significant failure of containment and detected downwind, much could be learned, including whether or not the North Korean device was based on plutonium or HEU.

Jungmin Kang is a science fellow at the Center for International Security and Cooperation (CISAC) at Stanford University. Frank N. von Hippel is professor of public and international affairs at Princeton University’s Program on Science and Global Security. Hui Zhang is a research associate in Harvard University’s Managing the Atom Project.


1. The complete text of the DNI statement was as follows: “Analysis of air samples collected on Oct. 11 detected radioactive debris confirming that North Korea conducted an underground nuclear explosion in the vicinity of P’unggye on Oct. 9. The explosion yield was less than a kiloton.”

2. Nevada Operations Office, U.S. Department of Energy, “Radiological Effluents Released From U.S. Continental Tests, 1961 through 1992,” DOE/NV-317, Rev. 1, 1996. With regard to the nomenclature of these isotopes, in Xenon-133m, for example, the number 133 designates the atomic number (neutrons plus protons). The suffix m indicates a long-lived (metastable) excited state of the nucleus.

3. Some correction for this effect could be made by measuring the iodine as well as xenon isotopes in the cloud. Martin Kalinowski’s unpublished article, “Characterization of Prompt and Delayed Atmospheric Radioactivity Releases From Underground Nuclear Tests at Nevada as a Function of Release Time,” does not identify evidence for such “fractionation” effects, but they still might be significant at the accuracy with which measurements would have to be made to distinguish between a plutonium and HEU bomb. Hui Zhang plans to carry though a more detailed analysis of this question.

Harold Smith Responds:

The Office of National Intelligence is, as it should be, cryptic in its official public announcements. The press is far less so. On Oct. 14, 2006, Mark Mazzetti of the New York Times reported on the North Korean nuclear test that a “senior intelligence official, who spoke on condition of anonymity, said that the results were still preliminary and that final analysis of the data would not be completed for several days.”[1] Three days later, Thom Shanker and David E. Sanger, experienced and respected New York Times reporters, wrote that “American intelligence agencies have concluded that North Korea’s test explosion last week was powered by plutonium that North Korea harvested from its small nuclear reactor, according to officials who have reviewed the results of atmospheric sampling since the blast.”[2] The question is, how did these anonymous officials reach their conclusion, which seems to be based solely on analysis of airborne radioactive debris, the skepticism of Kang, von Hippel, and Zhang notwithstanding?

The answer rests, in all probability, in the degree of accuracy required in the intelligence community (IC) versus that of the scientific community. It must be recognized that intelligence estimates are just what they purport to be, i.e., they are estimates. They are made on incomplete data under conditions quite unlike a laboratory, and most importantly, they have to be made in a timely manner. In this case, the IC could not wait for more data or further analysis; it had to decide within days whether the weapon was fueled by plutonium or uranium.

Based on activity during the past few years, plutonium seems far more likely. Indeed, Kang, von Hippel, and Zhang go to some pains to note that “we have no doubt that the material was plutonium.” But if the fuel were uranium, the government’s understanding of the nuclear capability of North Korea, already meager, would become even darker and more worrisome. Major and possibly counterproductive changes in strategy would be needed. Hence, it is entirely possible that, so long as the radio-chemical data was not inconsistent with a plutonium bomb, the IC felt comfortable in announcing, anonymously through the press, that the fuel was plutonium. Incontrovertible evidence, as Kang, von Hippel, and Zhang seem to want, would be nice, but it cannot be a requirement. It appears that nuclear forensics provided the necessary degree of comfort.

Harold Smith is a distinguished visiting scholar and professor at the Goldman School of Public Policy, University of California at Berkeley. He served as assistant to the secretary of defense for nuclear, chemical, and biological defense programs during the Clinton administration.


1. Mark Mazzetti, “Preliminary Samples Hint at North Korean Nuclear Test,” New York Times, October 14, 2006.

2. Thom Shanker and David E. Sanger, “North Korean Fuel Said to Be Plutonium,” New York Times, October 17, 2006.