A Technical Analysis of North Korea’s Oct. 9 Nuclear Test

Richard L. Garwin and Frank N. von Hippel

On Oct. 9, North Korea announced that it had carried out an underground nuclear test. In subsequent days, the apparent low yield of the device and initial lack of reports of detection of radioactivity from the test raised questions about whether North Korea had actually tested a nuclear device or if a test had failed.

One week later, the Office of the Director of National Intelligence issued a statement confirming the detection of radioactive debris and stating that North Korea had conducted a nuclear explosion with a yield of less than 1 kiloton.[1] Our analysis of the available public information supports this conclusion. We also judge that, while the test did not succeed as planned, North Korea might have been testing a lower-yield design than many commentators have assumed. This imperfect test may well lead North Korea to test again.

Our analysis concerns three questions: How powerful was the explosion? Was it a nuclear test? If nuclear, was the test successful?

How Powerful Was It?
North Korea reportedly informed China that it would be conducting a nuclear test, with a yield in the range of 4 kilotons.[2] Such an explosion in hard rock would produce a seismic event of magnitude 4.9.[3] By contrast, the U.S. Geological Service reported the explosion to have a seismic magnitude of 4.2.[4] South Korea’s state geology research center reported the magnitude as lower—between 3.58 and 3.7—and estimated a yield equivalent to 550 tons of TNT.[5] An uncertainty in seismic magnitude of 0.5 translates into an uncertainty in yield of about a factor of three. Compounding this uncertainty is that the relationship between seismic magnitude and yield depends upon the hardness of the rock in which the explosive is buried.[6]

Given all these uncertainties, it is not surprising that a range of yields has been reported. One authoritative estimate from Terry Wallace, a seismologist at Los Alamos National Laboratory, based on an unclassified analysis of open data, estimates a yield between 0.5 and 2 kilotons, with 90 percent confidence that the yield is less than 1 kiloton.[7] A second authority, Lynn R. Sykes of Columbia University estimates a yield of 0.4 kilotons, with 68 percent confidence that the yield is between 0.2 and 0.7 kilotons and a 95 percent probability that the yield is less than 1 kiloton.[8] One notable byproduct of the test is that it has demonstrated that university and other independent seismic detection systems, as well as those of governments and the International Monitoring System of the Vienna-based Comprehensive Test Ban Treaty Organization very effectively detect underground explosions in the sub-kiloton range.

Was it a Nuclear Test?
From the seismic data alone, the source might have been an explosion of a mixture of ammonium nitrate and fuel oil, ANFO, an inexpensive explosive used in mining all over the world. Five hundred tons of ANFO would fill the last 60 meters of a tunnel with a height and width of about 3 meters.

The radioactivity detected in the atmosphere of the region two days after the explosion could strengthen the evidence that it was indeed a nuclear test. No information has been made public about the nature of the radioactivity that has been detected. But, if a nuclear test occurred, particulate matter and gases might have been vented at the time of the test or radioactive gases might subsequently have seeped out through the cracks in the rocks above the explosion.

We focus here on two radioactive xenon isotopes, Xe-133 and Xe-135, with half-lives of about five days and 0.4 days respectively, that are often detected after underground tests. Xenon is chemically unreactive like helium and therefore does not plate out on the surfaces of cracks in the rock or get scrubbed out of the atmosphere by rain. We assume that the weapon was made of plutonium. North Korea is known to have enough plutonium to make at least several Nagasaki-type weapons although it is alleged also to have a clandestine uranium-enrichment program.[9] The ratio of the production of Xe-135 and Xe-133 in plutonium fission is known. Because the Xe-135 decays much more rapidly, the ratio of their concentrations in the plume provides a rough measure of the number of Xe-135 half lives and therefore the time since the test.[10]

It would take the fission of about 60 grams of plutonium to produce a yield of 1 kiloton. That much fission would produce about 2 grams each of Xe-133 and Xe-135. Because of their radioactivity, these xenon isotopes can be detected at a level of less than 100 atoms per cubic meter of air (one atom per 5x10²³).

Martin Kalinowski has provided us with a calculated trajectory of the first gas that might have been released by the test. By the end of the third day, he estimates that the plume would have traveled about 1,000 kilometers in a j-shaped track over the Sea of Japan.[11] At that point, the plume might be 1 kilometer high by 200 kilometers wide.  If the radioactive xenon produced by a 1 kiloton underground explosion were released to the atmosphere at a typical rate of 0.1 percent per day of the undecayed xenon, the concentration of xenon-133 in the plume would still be one hundred times above the detection limit. If the ratio of xenon-133 and xenon-135 concentrations was consistent with the time of the explosion, that would verify that it was nuclear.

Indeed, detection of Xe-133 alone after even a week or more could in itself confirm the nuclear nature of the explosion but its trajectory would have to be "backcast" to make sure that it was not due to leakage from reactors in South Korea or Japan. Much more could be learned if particles as well as gas leaked from the explosion, including its yield and whether it truly was a plutonium device.[12]

If Nuclear, Was It a Successful Test?
Much has been made of the apparently low yield of the North Korean test in comparison to the first U.S. plutonium explosive, the Nagasaki bomb, which had a yield of about 20 kilotons. In the Nagasaki bomb, tons of high explosive served to implode a solid subcritical sphere of plutonium to a higher density to make it supercritical. If Pyongyang tried to replicate the Nagasaki design, it is indeed likely that something went wrong.

Before the Nagasaki bomb was used in August 1945, J. Robert Oppenheimer, who directed the bomb-design effort at Los Alamos, wrote to General Leslie Groves, the overall head of the U.S. nuclear-weapon effort, that there was a 2 percent chance that the yield could be lower than 1 kiloton.[13] This would happen if a neutron started the chain reaction just when the plutonium first became critical, 10 millionths of a second before it reached its maximum supercriticality. The reason was that there was a 2 percent probability that the plutonium would produce a neutron spontaneously and start the chain reaction early. Other possibilities in the North Korean test are that extra neutrons might have been generated by the alpha particles from plutonium decays interacting with light-element impurities, or the neutron initiator might have been mistimed and fired too early or too late.

The fact that the weapon designers predicted a 4 kiloton yield suggests, however, that they were not aiming for the Nagasaki design. The Nagasaki bomb weighed about 4 tons, much more than could be lifted by any North Korean missile. Perhaps North Korea’s weapon designers tried to go directly to a weapon in the 500-1000 kilogram class that could reach South Korea on a Scud missile, or Japan on a Nodong missile—or the United States on a Scud launched from an offshore merchant ship.

A 4 kiloton or even a 1 kiloton explosive would still be a terrifying weapon. Recall that the 1995 Oklahoma City explosion involved only a few tons of ANFO. A 1 kiloton (1,000 ton TNT equivalent) bomb could kill people in an area of about one square mile and would partially destroy a much larger area.[14] Most of these deaths would be from fire or from the prompt nuclear radiation.

If, as the seismologists have concluded, the yield of the explosion was much less than the design yield, North Korea can have little faith in its nuclear weapon stockpile. It is likely that its weapons team will regroup. The North Korean government has already raised the possibility of a second test.[15]

Richard L. Garwin is IBM fellow emeritus at the IBM Research Center, Yorktown Heights, New York. Frank N. von Hippel is professor of public and international affairs, Princeton University.

ENDNOTES