On August 11, 1945 — just two days after the bombing of Nagasaki — the U.S. government issued a technical history of the Manhattan Project, written by Princeton physicist Henry DeWolf Smyth.1 The Smyth Report, as it came to be known (its official title was unpleasantly long), was meant to serve as the authoritative guide for what could be publicly said by Manhattan Project participants about the atomic bomb.
One of the areas that the Report was most sheepish about is how the actual charges of the atomic bombs — now called the "physics packages" — are designed. Implosion, the method used on the Trinity "Gadget" and the Nagasaki bomb ("Fat Man"), was ignored completely (and not declassified until 1951). Even the simple "gun-type" design used in the Hiroshima bomb, "Little Boy," was treated only obliquely:
Since estimates had been made of the speed that would bring together subcritical masses of U-235 rapidly enough to avoid predetonation, a good deal of thought had been given to practical methods of doing this. The obvious method of very rapidly assembling an atomic bomb was to shoot one part as a projectile in a gun against a second part as a target.2
In the early days, most people assumed that meant shooting two halves of a critical mass together, or, in more "real-looking" depictions, such as this very early one from the Austrian physicist Hans Thirring's Die Geschichte der Atombombe (1946), a small "projectile" being shot into a dense "target":
On Thirring's diagram,3 a "Phantasie" of "Details der Bombenkonstruktion" (you have to love the German here) based on the description in the Smyth Report, you can see that there is a projectile (P) which gets shot down an artillery barrel (R) by conventional explosives into the target (S), which is a larger amount of fissile material embedded in a tamper (T). The role of the tamper (which is discussed in the Smyth Report) is to reflect neutrons and hold together the fissioning mass a few milliseconds longer than it might otherwise be inclined. This allows for more fission reactions and more of an explosion.
So this is more or less how we've been talking about gun-type designs since 1945... until very recently. John Coster-Mullen, a trucker/photographer/bomb geek (and a friend of mine), dubbed "Atomic John," by the New Yorker in 2008, found, through some painstaking research, that this old story was wrong on one important detail.
The actual "Little Boy" bomb was not a small "projectile" being shot into the larger "target." It's a large "projectile" being shot into a smaller "target." That is, as John puts it, "Little Boy" was in fact a "girl":
Now half of you are saying "so what," the other half are saying "I already know this, I'm an atomic wonk," and the two of you who are not in that category (and are left out of the halves by rounding errors) are saying, "Cooooool."
There are some technical reasons why this arrangement is a better one, though I'm not the "technical boy" (to use a William Gibson phrase) who will expound on them, given my utter lack of claim to real technical authority.4 As an historian, what I find most compelling is the fact that I once I knew the above, I kept finding little bits and pieces to corroborate it sprinkled throughout the archive.
My favorite example: in 1964, William L. Laurence gave a long interview to the Columbia University Oral History Project, which is now available on microfilm at several major libraries. Laurence is a fascinating character who I'm sure to come back to in the future on here — he was a major pioneer in American science journalism, worked for The New York Times in the 1940s, and was the only member of the news media allowed to view the Manhattan Project. In this role he wrote basically all of the initial stories on the bomb that appeared after its use, which were distributed to all other newspapers in the country (and world) for reprinting. He got to ride along for the Nagasaki bombing, and won a Pulitzer for all of this atomic reporting. He's controversial and fascinating, but that's for another post.
Anyway, back to the interview. It's long, it's rambling, it's a few decades past the events in question. There are sections of his interview which are definitely suspicious from the standpoint of veracity. At one point he claims to have taught J. Robert Oppenheimer to be more eloquent. Riiight.
So mixed into this fact and likely-fantasy is a passage about the design of the Little Boy bomb:
[The bomb] contained the largest amount of uranium 235 assembled and concentrated up until that time. In fact, it contained as much as 66 kilograms of uranium 235, an enormous amount. ... It was oblong, cylindrical, and inside the quantity of uranium 235 in it was divided into two parts. The largest part was on one end attached to a gun mechanism, and the other part was on the other end. ... And one end was attached to a gun mechanism, and the other was in the front end of the cylinder, and so at the proper moment the gun mechanism would go off and fire the one part into the second part and thereby create a critical mass which would then explode automatically.
So this is pretty amazing for two reasons. One is that he gives a pretty specific number for the amount of the enriched uranium in the Little Boy bomb — 66 kg. Coster-Mullen has found sources stating it was around 64 kg; that's an incredibly close answer by Laurence for something that was definitely classified in 1964.
The other is that the description Laurence gives is the Coster-Mullen one: the "largest part" being shot into the "other part." Now this is the sort of detail that one might be inclined to miss, or to doubt, especially given the amount of total nonsense that inhabits the rest of Laurence's interview. But when you're looking for this sort of thing, it jumps out at you: here is Laurence independently (in the sense that Coster-Mullen had not seen this interview) confirming both the critical mass size (within 2 kg of error) and the unusual design mechanism. (Here also is Laurence committing a major security infraction, but such is life.)
So, I'm not entirely sure where I was going with all of the above, other than to give some illustration of 1. what sorts of wonkiness this blog will probably be covering (things too wonky-technical for my book); 2. the way in which the mixture of fact and "Phantasie" permeates all discussions of the bomb; 3. and the way that even a warmed-over topic like "how did the Little Boy bomb work?" can actually stand to have additional scrutiny brought to it, even in the publicly accessible sources. These are the sorts of things which make studying nuclear secrecy fun for me — there's just so much to delve into, to not be sure about, to independently correlate and confirm and doubt — above and beyond the more intellectual points.
- The paranoid pedant in me wants to point out that the date, August 11, is correct for the distribution date, whereas it is often quoted as August 12. In order to avoid any one newspaper getting the "scoop," the government requested that none report on it until the morning of the 12th, however. So either date is technically fine. Don't you feel better, knowing that? [↩]
- See §12.19, "Method of Assembly," in Chapter 12, "The Work on the Atomic Bomb." [↩]
- Those who are very into this bomb thing may recognize that this is the same image as the supposed "Nazi nuke" that made the rounds in 2005. Needless to say I am not super impressed with the claims that this was an actually working bomb and not just a visualization based on Thirring's book, which itself was clearly based on the Smyth Report. The fact that the "Nazi nuke" refers to the fissile material as "Plutonium," a name given to it in secret by Americans and only released after the bomb project was made project, makes it patently clear this is very much a postwar construction. [↩]
- Having the less massive piece be inside of the tamper cuts down on the possibility of pre-detonation, basically. You want your big, massive piece, to be outside of all of that reflection until the last possible moment, because you might inadvertently have a critical mass formed that will blow itself apart too early. Such is my understanding, anyway. I'm an historian, I don't run simulations! [↩]