The Third Core’s Revenge

by Alex Wellerstein, published August 16th, 2013

By the end of August 1945, there had been a total of three plutonium cores created in the entire world. Everyone knows about the first two. The first was put into the Gadget and detonated at Trinity in July 1945. The second was put into the Fat Man and detonated over Nagasaki in August 1945. The third, however, has been largely overlooked. The third core was the one that was destined to be the Third Shot dropped on Japan, had there been a Third Shot. Instead, it has a different story — but it was still not a peaceful one.1

The magnesium cases for the world's first three plutonium cores. Left: Herb Lehr at Trinity base camp with the Gadget core. Center: Luis Alvarez at Tinian with the Fat Man core. Right: The third core's case at Los Alamos, 1946.

The magnesium cases for the world’s first three plutonium cores. Left: Herb Lehr at Trinity base camp with the Gadget core, July 1945. Center: Luis Alvarez at Tinian with the Fat Man core, August 1945. Right: The third core’s case at Los Alamos, early 1946.

One of the questions I got from people regarding the “Why Nagasaki?” post I wrote last week was “When would the Third Shot really have been ready?” The reason for the question is that since the Third Shot was unlikely to have been ready by the time Hirohito announced Japan’s acquiescence to the American surrender demands (August 15), that satisfies the question of why another one wasn’t used. In a very practical sense, it does, but it ignores the fact that Truman actually put a “stop” on all further atomic bombings on August 10 — when the effect (if any) of the bombs on Japan’s high command was yet unknown. (He did not, it is worth noting, put a stop on firebombing: huge B-29 raids continued up until the surrender announcement.)

But still, it’s an interesting question to consider. There are two components to it: when did they think the third core would be ready, and when was it actually ready? On the first question, we know that on August 10, General Groves wrote to General Marshall that:

The next bomb of the implosion type had been scheduled to be ready for delivery on the target on the first good weather after 24 August 1945 . We have gained 4 days in manufacture and expect to ship from New Mexico on 12 or 13 August the final components. Providing there are no unforeseen difficulties in manufacture, in transportation to the theatre or after arrival in the theatre, the bomb should be ready for delivery on the first suitable weather after 17 or 18 August.2


It was on this document that Marshall scrawled, “It is not to be released on Japan without express authority from the President” — the Truman “stop” order. But we also know, from the Seeman-Hull document I discussed in an earlier post, that Marshall was still interested in the atomic production rate on Monday, August 13, 1945. At that time, Seeman claimed that:

Seeman: There’s one ready to be shipped now — waiting on order right now. […] The whole program is phased according to the best production. There is one of them that is ready to be shipped right now. The order was given Thursday [August 9?] and it should be ready the 19th.

Hull: If the order is given now, when can it be ready?

Seeman: Thursday [August 16] would be its readiness; the 19th it would be dropped.

Hull: In other words, three or four day advance notice before it can be shipped, and six days after that when it can be dropped.

So that’s a pretty interesting conversation — it tells us that the core was in some kind of almost-finished state by August 13. In a 2012 interview, physicist Lawrence Litz told Alexandra Levy of the Atomic Heritage Foundation that:

Levy: What was—how did—do you remember working on casting the plutonium for the third bomb?

Litz: The particular day that remembers—that remains in my memory was the day that we cast the plutonium for the third bomb because we weren’t sure that the Japanese would surrender even after the second bomb was dropped. We had to cast the atmospheres for the third, and because time was short we had to cast the two hemispheres at the same time. But it was dangerous to cast them in the same laboratory at the same time so we set up two adjacent laboratories with the high vacuum apparatus and the—so we could cast one hemisphere in each one of the two labs.

Levy: How long did that take to cast?

Litz: About twenty-four hours and we had to work straight through.

Which gives some indication of the tenor of the day, and the fact that Truman’s “stop” order didn’t mean that they weren’t expecting to potentially keep atomic bombing. (As does the Seeman-Hull conversation.)

How much plutonium was on hand in August 1945? I’ve been hunting around for anything that would give me some hard numbers on this, and finally, basically when I’d given up on the effort, I was surprised to stumbled across a document that did:

1945-08-30 - Los Alamos plutonium inventory

“49 Interim Processing Program No. 24,” dated August 30, 1945, indicates that by that date that Los Alamos had, by their assessment, received 26.136 kg of plutonium from Hanford.3  Figuring out what was done with all of that requires a little decoding of the terminology. 12.292 kg of the material is listed as having been transferred to the US Army with the notation “HS-1, 2, 3, 4; R-1” after it. I haven’t seen this notation before, but I think it’s almost certain that “HS” means “hemisphere,” i.e. half of a sphere of plutonium. So two full spheres worth were transferred to the Army and were at that time “non-usable,” along with “R-1.” R-1 is almost certainly an “anti-jet” ring developed for use in the Fat Man core (and not present in Trinity’s core).4 So HS-1+HS-2 were the Trinity core components, and HS-3+HS-4+R-1 was the Fat Man core. The first two cores were “non-usable” because they had been detonated.

So we can see from the document that HS-5, HS-6, and R-2 had already been cast and were in the hands of Quality Control at the lab (QC). HS-7 and R-3 had been already cast by then, but still needed hot pressing and nickel coating. HS-8 was scheduled to be pressed on August 31, and finished by September 5. Which is the finest-grain look at the early nuclear production schedule that I’ve ever seen. (And as you can tell I’m quite proud of myself for finding it and deciphering it!)

But the story of the third core doesn’t end there. 

The core was cast sometime around August 13th, but still likely needed to be pressed and coated, ergo the need to take until August 16th to finalize. By August 15th, it became clear that it wasn’t going to be needed in the war. So it was kept at Los Alamos.

A mockup of the third core's experimental setup, August 21, 1945. (Source: Los Alamos)

A mockup of the third core’s experimental setup, August 21, 1945. (Source: Los Alamos)

What it was doing between August 15th and August 21st, I don’t know. But I do know that on August 21st it was being used for critical mass experiments — “tickling the dragon’s tail.” The experiments in question involved surrounding a full 6.2 kg core with tungsten carbide, getting information about the effect that different tamper arrangements had on criticality. (The tamper reflects neutrons back into the core, thus increasing the overall neutron economy and thus lowering the effective critical mass.)

The experimenter in question was 24-year-old physicist Harry Daghlian, Jr. To quote from a report on the experiment:

[Daghlian] was carrying one brick [of tungsten carbide] in his left hand over the assembly, to place it in the center of the fifth layer. While he had this brick suspended over the assembly, he noticed (from the instruments) that the addition of this brick would have made the assembly supercritical if placed on top of the assembly. Having realized this, he was withdrawing his left hand and the brick from over the assembly and while doing so the brick slipped out of his hand and fell immediately onto the center of the assembly. Knowing that this brick would made the assembly dangerous, he instinctively and immediately pushed this brick off the assembly with his right hand. While doing this, he stated that he felt a tingling sensation in his right hand and at the same time noticed a blue glow surrounding the assembly, the depth of the blue glow being estimated to be about two inches.5

Daghlian was estimated to have received a 510 rem dose of ionizing radiation — a usually lethal dose. He died after an agonizing month. This, incidentally, appears to have been why at the time of the August 30 audit, the core was in Quality Control: they were checking to make sure it had not undergone any “dimensional changes” as a result.

One might think that someone involved with the investigation of the Daghlian accident would be especially cautious around using such a core in further critical mass experiments, even if only for superstitious reasons.

Re-creation of Slotin's fatal experiment with the third core. (Source: Los Alamos)

Re-creation of Slotin’s fatal experiment with the third core. (Source: Los Alamos)

But exactly 9 months later, one of the co-authors of the above-cited report, Louis Slotin, would himself receive a lethal radiation dose from the exact same core in the process of yet another (different) critical mass experiment. Slotin knew the experiment in question was dangerous, and had been told by Enrico Fermi that he would be “dead within a year” if he continued to work with such bravado. Like Daghlian, his hand faltered at a literally critical juncture: he was holding a neutron reflector above the core with a screw driver when his fatal slip occurred, lowering the reflector just a fraction of an inch, releasing a stream of neutrons and the characteristic blue glow. Slotin died 9 days later.

The third core, by now nicknamed the “demon core” for having taken two lives, would not go out with a whimper. By some accounts, it found its final disposition in the first postwar nuclear test, shot “Able” of Operation Crossroads, on July 1, 1946,  just under a year after it had been first cast, in that all-night session, in the closing days of World War II. (UPDATED BELOW) Encased in a “Fat Man” assembly with “GILDA” stenciled on its hull, it was finally dropped from a B-29, as it was originally intended to be, and it detonated over a fleet of empty ships in the Bikini atoll, with a yield of 21 kilotons. Alas, the journalists who saw it, with perhaps higher expectations for their first atomic bomb test, incorrectly dubbed it a flop.

The final use of the third core: the Crossroads "Able" shot, July 1, 1946.

The final use of the third core: the Crossroads “Able” shot, July 1, 1946.

That a single plutonium core could go through so much may seem remarkable. But it is a reflection of a time when such cores were extremely rare commodities. And so a single core could simultaneously be the one originally destined for the “third shot,” and also be the subject of two fatal criticality accidents, and also still be the first core consumed by postwar nuclear testing. It is a potent reminder of how paltry the American nuclear arsenal once was — when there were less than a dozen pieces of cores, much less cores themselves.

UPDATE: The third core was not used at Crossroads, after all! Learn what happened to it, here.

  1. Since a few people have gotten confused, I think I should say somewhere explicitly: the Hiroshima bomb, Little Boy, used a 64 kg highly-enriched uranium core. I’m only talking about plutonium here, in part because it was only plutonium cores that were being manufactured at this point, since the Little Boy design was considered more or less instantly obsolete. []
  2. Leslie R. Groves to George C. Marshall (10 August 1945), copy in the Nuclear Testing Archive, document NV0137881. []
  3. C.S. Garner, “49 Interim Processing Program No. 24,” (30 August 1945), DOE OpenNet Document ALLAOSTI126018. It is interesting, as well, that the Hanford (W) and Los Alamos (Y) assays were off by 1.376 kg, which is quite a lot in this context (22% of a bomb core, or 44% of a single hemisphere). There are indications in the files that they did quite a lot of sniffing around trying to figure out what each site was doing that led to these different assessments. The problem of Material Unaccounted For never really goes away, but it’s interesting that it shows up this early in the game. []
  4. I discussed the fact that the Trinity and Nagasaki cores were slightly different in a very old blog post; Trinity was just two hemispheres, whereas Fat Man also included the ring. []
  5. Paul Aebersold, Louis Hempelmann, and Louis Slotin, “Report on Accident of August 21, 1945 at Omega Site,” (26 August 1945), LAMD-120, copy reprinted in John Coster-Mullen, Atom bombs: The Top Secret inside story of Little Boy and Fat Man, rev. 2007. []

42 Responses to “The Third Core’s Revenge”

  1. burl grey says:

    Breathtaking reading!

    • Blake says:

      Seconded, this is absolutely fascinating. How the hell is he finding this incredibly obscure shit?! lol. I can’t remember if I already asked somewhere, but I wonder if there is a paypal link or something we can use to donate to the blog, it’s definitely THAT good.

      • AZGuy says:

        Check the Atomic Testing Museum in Las Vegas, and the National Nuclear Archives in the same building. They have all the declassified records on nuclear research in America for research purposes. It might be where he found these.

  2. phuzz says:

    I suppose if I’d paid attention to the title I’d have worked it out sooner, but I’d never realised that the demon core was the third one manufactured.
    Being able to see Chrenkov radiation (the blue glow) with the naked eye, is generally a bad thing.

    • I hadn’t put two-and-two together until very recently, myself.

      Also, interesting fact, the blue flash from criticality accidents is not Cherenkov radiation. Cherenkov radiation is caused by electrons moving quickly through a slow medium (like water), whereas the blue flash is caused by the excitation and subsequent spectral emission of the air. The fact that they are both blue is just a funny coincidence. Either way, though, neither are anything you want to see in your immediate vicinity!

      • Old Salt says:

        And in addition ~6 cc’s of humor in each eyeball exposed to a some positive charged particle flux from a very small percentage of ternary fission, possibly providing an internal blue glow from within the eyeball? “Ugly bags of mostly water” have downsides!

      • Blake says:

        Disturbingly, this may not have even been Slotin’s first time seeing or touching that enigmatic and beautiful Cherenkov glow. A few months before his fatal misadventure he swam to the bottom of a borated water shielding pool on top of the Clinton pile at ORNL WHILE IT WAS OPERATING to repair one of his experiments that he didn’t want to wait another 2 days until shutdown to fix. He took off his film badge before doing it of course, so it was only roughly estimated later that he likely blasted himself with an obscene ~>100R of exposure. Fermi was bang on in his warning that Slotin’s gross recklessness would soon see him dead.

      • kme says:

        It’s quite safe to observe the Cherenkov radiation in swimming pool type nuclear reactor.

      • SeanB says:

        We see something similar with synchrotron radiation. ~20 years ago most monochromators were He-filled rather than in vacuum. The incident (broad spectrum) X-ray radiation would hit the silicon crystal and most of the beam would be absorbed, only a single wavelength would be transmitted. Much of the absorbed beam would be re-emitted as photo-electrons and it is these energetic electrons which caused the He to glow blue. So I suspect it is the beta decays that are the source of the blue glow that was seen. In vacuum you don’t see any blue glow, nor do you see it as the x rays pass through the He before hitting the monochromator crystal. These electrons are unlikely to be the reason those guys picked up hundreds of REM. It is likely the neutrons that killed them.

  3. Fascinating and frightening! We expect weapons of war to kill people, but to kill their creators?

    Congratulations on uncovering new information from so long ago and writing about it so informatively.

  4. Ben says:

    Regarding the first accident with Daghlian, what would have happened had he not removed the brick?

    • Probably heated until it knocked itself apart, like the Godiva device did. I don’t think it was at much risk of exploding as it was just barely supercritical and it didn’t take much to stop the reaction. But I don’t really know.

    • We know it would not create a nuclear explosion as that is only possible by implosion. Presumably, the core would continue to emit radiation at the interface with the brick and eventually enough heat to melt either or both, which could quench the reaction.

      • Ben says:

        It’s been a while since I’ve seen the movie with Newman and Cusack, but I remember them kind of combining the two criticality accidents into one dramatic representation. They make it out like Cusack saved his comrades by jumping in and scattering the assembly. It sounds like that is not the case? If everyone ran they might have a really ‘dirty’ building but nothing else?

        • In both the Daghlian and Slotin accidents, they quickly tried to disassemble the device and in the process may have saved the other people in the room from further exposure. (There was an SED guard with Daghlian not far from his setup, and there were many people in the Slotin accident room.) So it’s not entirely fictitious to have the accident-creator being brave or something like that.

          • Puncheex says:

            It is thought that in Slotin’s case, the beryllium reflectors disassembled themselves after a few milliseconds of intense heating distorted the metal’s shape, but if Slotin hadn’t thrown off the top reflector, it would have settled back down after cooling a bit, and the accident would have repeated. That makes his gesture no less heroic by intent.

        • The accident was caused by ‘dropping’ the brick, which fell at the gravitational constant of 32.2 ft/sec^2 … if I have my high school physics right. That’s a lot faster than the human nervous system, much less the human musculoskeletal system. My point … it was not humanly possible to save anyone, much less himself by reflex action … he was a ‘fatality’ within milliseconds, but it took a month for Daghlian to die and nine days for Slotin.

          • Ben says:

            In the back of my mind I remember reading something about one of the accidents where Slotin/Daghlian (or others in the room) reported a ‘metallic’ taste in their mouths at the time of the accident. Along the same lines of seeing the blue flash in your immediate vicinity it’s probably not good to taste metal either.

            I can only imagine what the two men thought right after the accidents. It must have been strange to have had no immediate outward trauma while at the same time knowing the end is coming.

          • Sven Geier says:

            I’d be tempted to ascribe that “Metallic taste” (possibly “metallic aroma”) to ionized air: in high-energy discharges (like immediately after lightning strikes, but also in industrial settings) people describe a metallic taste and a “smell of sulphur” (the latter really being ozone).

          • Cthippo says:

            The smell of ozone has been reported fairly frequently by victims of high gamma dose exposure, including several who have accidentally walked into operating medical irradiators. Search for accident reports on the IAEA website for a whole series of these reports.

  5. Wesley says:

    Can you help me find a first-hand reference that confirms that Crossroads Able used the the same plutonium as in the criticality experiments. I tried once and was led down a circuitous path that included the book Under The Cloud (which gave no reference to its statement). I’ve always wondered why the Pu could not have been that used in Baker or the third planned shot or the few left in the stockpile at that time. If I recall correctly we had 6 weapons in the stockpile at the end of 1945 (Crossroads took place in 1946).

    • Wesley — interesting question. I haven’t found any direct documentation of that, though I’ve seen it asserted that the demon core was detonated at Able many times.

      Poking around, the only potentially relevant document I’ve found is this one, from November-December 1946, which says that “HS-5, HS-6 and R-2 passed a second final inspection after HS-5 was stripped, recleaned and recoated.” If HS-5, HS-6, and R-2 are the third core (as they appear to be in the document from August 1945), then that would indicate that they were not used in Crossroads.

      But that’s the only little snippet of potentially relevant information that I’ve seen, and it hinges on the idea that they didn’t re-label the core inventories between August 1945 and December 1946.

      • John Coster-Mullen says:

        My guess here is that HS-5 and HS-6 are the individual Pu hemispheres and R-2 is the Pu jet ring that fit in between the hemispheres. If true, then this core was not used at Crossroads.

        Following their numbering sequence, HS-1 and 2 were used at Trinity, HS-3 and 4 (along with R-1) were used at Nagasaki. A jet ring was not used at Trinity. They used pieces of gold foil instead and fabricated the jet ring (with a triangular cross-section) for use “in the field” because the gold foil method used in the McDonald Ranch House assembly was too tedious and complicated.

  6. kme says:

    The third core, by now nicknamed the “demon core” for having taken two lives…

    That seems a bit rough, given how many lives the second core took!

  7. Jonathan Walton says:

    There was a shop in Los Alamos that was sort of a surplus store/warehouse of laboratory discards/castoffs. When I went in there in the late 90’s they had a hemispherical die which was used to mate the explosives (as I recall) around the core. It would have held about a soccer ball size mass but it looked sort of like a metal font or fountain.

  8. […] Wellerstein, The Nuclear Secrecy Blog, The Third Core’s Revenge, here. […]

  9. Stephen White says:

    The story of the “demon core” appears in the the docudrama Fat Man and Little Boy, with John Cusack playing a fictional composite of Slotin and Daghlian.

  10. LegoDoug says:

    Very interesting, indeed.

    If I remember correctly, Richard Feynman mentions one of these critical mass incidents in far less detail, either in Surely You’re Joking, Mr. Feynman! or What Do You Care What Other People Think?

    I’d also heard from time to time, incorrectly I see, that we did not have the plutonium for a third nuclear bomb after Nagasaki. I am glad to know the rest of the story.

  11. Chew says:

    The Caustic Soda podcast did a hilarious bit on the Demon Core in this episode:

  12. Dcotor TOC says:

    Fascinating bit of history, and a great introduction to the blog.

  13. Stan Norris says:

    There is a 1979 Los Alamos Report (LA-UR-79-2802) titled “What has Happened to the Survivors of the Early Los Alamos Nuclear Accidents.”

  14. […] the story on the simulation. I found a fascinating article about the third nuclear core built for Japan in World War II. Since the second nuke did the trick, […]

  15. John Coster-Mullen says:

    The third Fat Man bomb (Unit F-32) was already on Tinian. It had been delivered along with F-31 (used at Nagasaki) and F-33 that had been dropped as a non-nuclear test off Tinian on Aug. 8. Raemer Schreiber was scheduled to deliver the F-32 core when he was stopped before he left Los Alamos.

    Whether true or not, one source told me years ago that the core mentioned here was actually called “Rufus” and was slightly “over” criticality specifications. Any tamper material experiments done with this core would therefore make an actual core use a little safer.

  16. […] Gadget sphere itself, and the people standing around (including Louis Slotin, who would later be killed by playing with a core). It’s kind of surprising it was declassified, since the length of the tamper plug is the […]

  17. […] project for those who made it. Occasionally there were more exotic threats. Criticality accidents took the lives of two scientists in the immediate postwar, as is well known. Concerns about criticality excursions at the plants […]

  18. […] to be as it was played out — e.g., using both bombs as soon as possible, in early August. A third plutonium bomb would have been ready by August 17th or 18th (they originally thought the 24th, but it got pushed up), so one could imagine a situation in […]

  19. […] of Louis Slotin’s fatal experiment with the third plutonium core. The problem wasn’t the mass of the core, it was that Slotin inadvertently changed the state […]