When Morgan returned to work the following Monday, he discovered that Slotin had stripped down to his shorts, dived into the pool, and made his adjustments. Morgan was appalled. Slotin was reassigned to Los Alamos, where daring was better appreciated, in December 1944. He quickly earned respect for a natural ability to assemble the complicated implosion bomb without excessive worrying and hand-wringing. He expertly put together the bomb core for the Trinity test in New Mexico in July 1945. His unofficial title was Chief Armorer of the United States. The only reason he was absent on Tinian Island when the Fat Man was assembled was his lack of U.S. citizenship.
Slotin was shocked and saddened when Daghlian, his assistant and fellow dragon tickler, died in the criticality accident, and he spent days at his bedside in the hospital. This tragedy, however, did not affect his supreme confidence. He brushed aside advice that he should automate the critical assembly experiments, even when the very wise Fermi warned him that he wouldn’t last a year if he kept doing that experiment. The central problem pointed out by Daghlian’s death was approaching criticality from the top, where gravity could accidentally complete the operation. It would make more sense to assemble from the bottom. If anything was dropped, it would fall away from the plutonium sphere instead of into it. Slotin discounted the advice as an unnecessary complication.
The next atomic bomb explosion was to be a test in the middle of the Pacific Ocean at Bikini Atoll, designed to demonstrate that a navy flotilla could survive a nuclear attack and proving that the water-borne armaments had not been made obsolete by this recent innovation. The date for the Able shot in Operation Crossroads was set for July 1, 1946. The implosion bomb was under constant improvement, and the WC tamper had been replaced by a beryllium tamper. It was machined into a pair of concentric shells, 9.0 and 13.0 inches in diameter, and split into hemispheres to fit around the plutonium bomb core. The beryllium would act as a secondary neutron source during the explosion, hopefully increasing the number of fissions as the core destructed. A hole was bored into the top hemisphere so that the “initiator” modulated neutron source could be inserted into the core without dismantling the entire bomb. The criticality experiment for this revised tamper design was moved to a new building in Pajarito Canyon, and it would be performed using the same ball of delta-phase plutonium that had killed Daghlian.
On May 21, 1946, Slotin was training his replacement, Alvin C. Graves, who had a Ph.D. in physics from the University of Chicago. Slotin had grown weary of the bomb work, and was planning to bail out and go work in biochemistry back east. This would be his last bomb core.
It was about 3:15 in the afternoon. The experiment was set up on the low assembly bench, with the bomb set up near the edge and the 5 millicurie Ra-Be neutron source placed a few inches in front of it.[48] Radiation detectors of several types were set up on and near the bench and warmed up, giving continuous recordings and audible clicks. There were seven men in the room, which was unusual for a criticality experiment, but this one was informal and was not scheduled. Two men had been working on initiator tests on a bench on the east side of the room, and the sensitivity of the required radiation counts had delayed them several times as the background counts were perturbed by tests outside the building. Slotin’s demo for Graves would also interrupt them, but it would be interesting to see him do the now famously dangerous criticality test. The SED guard was present, as were two other scientists, and they were all fascinated by watching the skilled Armorer at work. Three were standing directly in front of the bench. The room was brightly lit with overhead fluorescents and low sunlight through the windows.
The formal test called for wooden spacers to hold the top nine-inch tamper hemisphere off the bottom hemisphere, and it was to be gradually lowered onto the core by changing out the spacers, one at a time, with smaller ones until the assembly was very close to criticality. Both the 13-inch and the 9-inch tamper hemispheres were installed only on the bottom of the assembly. The tamper pieces would then be sent back to the shop to have some metal removed, and the test would be done over until the assembled bomb was stable and very near the critical condition.
Slotin discarded the spacers and used a big-bladed screwdriver instead. With the blade under the lip of the tamper, he could lever it up and down, impressing Graves and his audience by making the neutron count rate zoom in and out on the loudspeaker. Graves was close behind him, looking over Slotin’s right shoulder. Slotin’s left thumb was through the access hole on top of the tamper, with his fingers on the curved side, adding to the downward tension of the movable tamper-half. He pulled the screwdriver handle up, increasing the angle between the bottom hemisphere and the straight blade, with the top hemisphere riding up, increasing the gap and making the neutron count rate fall precipitously. At an angle of 45 degrees, the screwdriver arrangement became precarious, as the side thrust, pushing the screwdriver outward from the gap, equaled the downward thrust holding the screwdriver down. Ever upward Slotin angled the tool. Beyond 45 degrees, the outward thrust overcame the downward thrust, and the screwdriver suddenly escaped the gap.
Bang.
The top tamper fell squarely on the bomb assembly, and prompt criticality was achieved instantly.[49] The blue flash lit up the entire room, as the neutron counters, ticking merrily along, suddenly jammed and went quiet. Slotin, on pure instinct, jerked the tamper off the assembly and dropped it on the floor. He could feel the tingling in his left hand and he could taste the radiation on his tongue. It had happened again, and there was no ignorance at work here. Familiarity to the point of nonchalance had just claimed another victim.
Slotin had a body dose of 2,100 rem of mixed radiation, or twice the dose of guaranteed lethality, and he died the same way Daghlian had, only faster, nine days later. The same radiation pulse hit Alvin Graves, standing an inch from Slotin, but he was partly shielded by the Armorer’s body. He stayed in the hospital a few days and was released. The other men in the room showed minimal effects from the incident.
The Crossroads tests went on as planned, with the Able shot using the bomb core that had killed two scientists. The bomb, dropped from a B-29, was affectionately named Gilda, and had a picture of Rita Hayworth painted on the side. The yield was 23 kilotons, or 3 kilotons more than the device dropped on Nagasaki a year earlier with a uranium tamper. The target flotilla consisted of 95 vessels of all types, from a captured Japanese battleship to a floating drydock.[50] All were sunk, lost, damaged beyond repair, or made dangerously radioactive except one, the U.S. submarine Dentuna, which was refurbished and returned briefly to naval service. The ships were manned by 57 guinea pigs, 109 mice, 146 pigs, 176 goats, and 3,030 white mice. Some lived through the air blast and the radiation pulse, with the most famous survivor being Pig 311, who was found swimming in Bikini lagoon after it stopped raining battleship fragments. He lived out his life at the Smithsonian Zoological Park in Washington, D.C., on a government pension.
There would never be another manual bomb assembly experiment, anywhere or any time. There was still a need to test-assemble the core parts to find unwanted critical conditions, and even an application of a chain-reacting naked plutonium core, to produce the specific radiation spectrum of an atomic bomb explosion. All further work was done at a distance of a quarter of a mile, using remote controls, television cameras, and a quick shutdown capability. The practice of bringing very small, bare metal reactors to the power-production point was still an extremely ticklish, sensitive action, but at least nobody could get hurt. That was the intent, at least, but fissionable materials always seemed capable of finding a flaw in the best intentions. All the remote-controlled assemblies in the United States were named “Godiva.”[51]
48
Declassified documents differ on the type of neutron source used. One says it was a radium-beryllium source, and one says it was plutonium-beryllium. Usually not mentioned was an extremely active 30 curie polonium-beryllium source, three months old, located about seven feet east-northeast of the assembly table. These seem minor points, but one strives for as much accuracy as possible.
49
“Prompt criticality” is an important term. It means that a state of at least break-even chain reaction was reached using only promptly available neutrons. There was no need to wait for the delayed fission neutrons, which could take seconds, to get a load sufficient to declare the assembly critical. The additional reactivity needed to go from delayed to prompt criticality is exactly one dollar. The practice of expressing reactivity in dollars and cents, still in use today, was coined by Dr. Louis A. Slotin.
50
The Japanese battleship was the
51
Not really. At the Los Alamos Lab there were also the Topsy and the Jezebel assemblies. These specialized devices may have been better designed, as no unplanned criticalities were ever reported for them. The French were conducting similar experiments at about the same time, first at Saclay with the PROSERPINE and ALECTO assemblies and then after 1961 at the Valduc Research Centre using the CASTOR and POLLUX machines (rig B and rig D). No accidents were recorded.