This week's document is one that most nuclear wonks have seen before: the Summary Report of the Nth Country Experiment, produced by the Lawrence Livermore laboratory in 1967. It made a big splash when it was declassified in 2003, for good reason. Here was an official government study, from over 30 years ago, which said that there were essentially no secrets left when it came to designing nuclear weapons.1
The report summarizes the results of a 1964 "experiment," in which Livermore hired two physics postdocs and had them try to come up with "a credible weapon design" based only on information in the public domain and computer support. A third postdoc was added a year later. The experiment ended in April 1967.
From a modern perspective, this is fascinating stuff. If three postdocs can design a nuclear weapon, then what's to stop a terrorist? What's the value of secrecy? Isn't it amazing they were worried about this stuff almost 50 years ago?
But I think there are some more things to say about this.
First off, there's one big, huge, hulking elephant in the room. Were they actually able to design a credible nuclear weapon? That... turns out to be not clear at all. The report is heavily redacted, and all of the parts that ought to clarify this very basic question are missing. It's clear the postdocs though they designed a credible nuclear weapon -- one that could be deemed credible even without an kind of actual test. They chose a more difficult task -- designing an implosion bomb -- just because they thought it would be more interesting. From the outlines it seems they basically tried to replicate the "Fat Man" design, plutonium-beryllium initiator and all.
Of the evaluation, we have the following to go on from the "Critique of the Nth Country Weapon Design" by F.S. Eby and L.S. Germain (page 28 in the PDF above):
- "As the reader will discover below, the detailed LRL design calculations, using codes unavailable to the Nth Country physicists, disagree with both of these numbers." This is, of course, an isolated sentence in a sea of redactions.
- "It was impossible for us to detail the Nth Country predictions from the data given in the sections on Final Design and Test of the Nth Country Experiment report. [lots of redactions] They give no estimate of the magnitude of the latter effect on the final yield and, in fact, do not really cite reasons for their belief that the alpha decreases too rapidly. They correctly observe that they have very little firm information about the criticality of their system. [more redactions] In light of this extreme sensitivity, it would seem that confidence int he expected yield is unwarranted."
- "Another point which would appear to call for conservatism in the prediction of the yield of the test device is [redacted]."
- "In summary, the Nth Country designers arrived [redacted] The authors do not give any detailed reasons for the discrepancy. [redacted]"
That's not much to go on, but it is all along a similar theme: there were a lot of unknown unknowns for the postdocs, and their confidence that their weapon would work as planned is probably unwarranted. At no place in the Summary Report does it imply that anyone but the postdocs thought the design would work as expected.
This is, incidentally, exactly what I would expect from a homemade nuclear weapon design. The hardest part about designing a nuclear weapon in the post-Manhattan Project era, from what I gather, is having any confidence that it will work the way you expect it to. American and Soviet confidence in their own weapons is the product of thousands of nuclear test explosions, from which they calibrated their equations and their understanding of the "art" of such design. Simulating a specific nuclear weapon design from first principles, though, is not really in the cards, even today, much less in 1964. (This was the same criticism that was made of the "homemade" designs of John Aristotle Phillips and Dimitri Rotow in the 1970s.)
Ah, but does this matter? Would their design likely produce some yield even if it wasn't maximally efficient? This gets us into a slightly more modern debate than what was going on in 1964. In 1964, they were concerned about the Nth Country: it's a study about proliferation, not terrorism. For proliferation to matter, they reasoned, it would have to be a "militarily significant" yield -- a dozen kilotons or so. That's the concern, and it runs explicitly throughout the decisions made by the postdocs about their designs and fissile material choices.
For a terrorist, though, a sub-kiloton yield could accomplish enough mayhem and death to suit anyone, if it was placed in the right place. That was Ted Taylor's argument when he started the real alarm about nuclear terrorism (with his and Manson Willrich's Nuclear Theft: Risks and Safeguards, from 1974, released on the heels of John McPhee's classic portrait of Taylor, The Curve of Binding Energy). But if that was the goal, then you probably be worried about a complicated implosion design from scratch anyway (much less many of the other design decisions that the postdocs made), so the Nth Country Experiment is probably useless on that point as well.
To me, this all adds up to one big conclusion: citing the Nth Country Experiment as evidence that nuclear weapons are easy to design these days seems unwarranted, if not actually misleading. We don't know for sure that the Nth Country Experiment design was a flop, but we have no evidence that it was a success.
The closest thing we have is a recollection of a comment from decades earlier that was reported in The Guardian:
Finally, after a valedictory presentation at Livermore attended by a grumpy Edward Teller, they were pulled aside by a senior researcher, Jim Frank. "Jim said, 'I bet you guys want to know how it turned out,'" Dobson recalls. "We said yes. And he told us that if it had been constructed, it would have made a pretty impressive bang." How impressive, they wanted to know. "On the same order of magnitude as Hiroshima," Frank replied.2
(Poor Edward Teller -- always grumpy!)
"On the same order of magnitude as Hiroshima" is a nice, fairly empty phrase if you don't know the units being used (much less if it is accurate). A kiloton? 10 kilotons? More? Less? I don't know.
My suspicion is that it was probably not more than a few kilotons, if anything. Why? Not because I don't think you could design a crude nuke without too much difficulty -- rather, because I don't think three physics postdocs combine enough of the requisite skills and experience to design an implosion bomb from scratch that they would have much confidence in working. (Three more advanced physicists with experience in relevant areas, and some engineering and chemistry acumen, could be a different story.) This is just my hunch, though.
(I won't even bother pointing out that "designing" an atomic bomb on paper is, of course, nothing doing compared to actually producing the fissile material, casting explosives, fabricating the right shapes of things, assembling the whole device, all the while not killing yourself in the process. Headlines aside, these guys did not build an atomic bomb in any sense of the term "build," which I think most thoughtful nuclear observers realize.)
- Source: W.J. Frank, ed., "Summary Report of the Nth Country Experiment," UCRL-50249 (March 1967), via the National Security Archive. [↩]
- Oliver Burkemann, "How Two Students Built an A-Bomb," The Guardian (23 June 2003). [↩]