The Sphere
It was 3.5 inches in diameter and weighed 14 pounds — a sphere of plutonium-gallium alloy, machined to extraordinary smoothness at the Los Alamos weapons laboratory in New Mexico. Designated as a weapons core, it was originally intended for use in a third atomic bomb to be dropped on Japan in August 1945. The bomb it was destined for — nicknamed "Rufus" — was being assembled when Japan surrendered on August 15 [1]. With no target remaining, the core was retained at Los Alamos for criticality experiments: tests designed to measure how close a mass of fissile material could be brought to a self-sustaining nuclear chain reaction.
Within a year, this sphere would kill two scientists in nearly identical accidents and acquire a name whispered through the hallways of America's most secret laboratory: the Demon Core.
What Criticality Means
A nuclear chain reaction occurs when enough fissile material is assembled in the right geometry for neutrons released by one fission event to trigger additional fissions faster than they escape or are absorbed. The exact amount needed — the "critical mass" — depends on the material's shape, density, and what surrounds it [2]. Reflectors like tungsten carbide or beryllium bounce escaping neutrons back into the core, reducing the critical mass needed.
Criticality experiments at Los Alamos were designed to determine exactly where this threshold lay for various configurations. Scientists would incrementally bring a core closer to criticality — adding reflector material brick by brick, or closing reflector shells around the core millimeter by millimeter — while monitoring neutron counters [1]. The goal was to approach the edge as closely as possible without crossing it.
The margin for error was essentially zero. A chain reaction, once started, released lethal radiation in microseconds — far faster than any human could react. These experiments were, in a very real sense, gambling with physics [3].
Harry Daghlian: August 21, 1945
On the evening of August 21, 1945 — just six days after Japan's surrender — 24-year-old physicist Harry Daghlian was working alone in a laboratory at the Omega Site, a remote canyon complex at Los Alamos. He was conducting a criticality experiment: stacking tungsten carbide bricks around the plutonium core to measure how much reflector material was needed to bring it to criticality [2].
The procedure itself was authorized, but the timing and conditions were not. Daghlian was working at night, alone, in violation of safety rules that required a second scientist present for all criticality work. A security guard, Private Robert Hemmerly, sat in the room reading a newspaper but had no scientific training and could not have intervened meaningfully [1].
As Daghlian lowered the final tungsten brick — the one his calculations suggested would bring the assembly to the edge of criticality — his hand slipped. The brick fell directly onto the assembly, pushing the core past critical. A blue flash of Cherenkov radiation filled the room. Daghlian felt a wave of heat wash over his body [3].
He instinctively knocked the brick away with his right hand, stopping the chain reaction. But he had already received a lethal dose of radiation — estimated at 510 rem over his entire body, with his right hand absorbing approximately 20,000 rem [2]. He walked out of the lab under his own power, reported to the medical facility, and told the duty officer what had happened.
Over the next 25 days, his body disintegrated cell by cell. His right hand swelled grotesquely, blistered, and began to slough tissue. His white blood cell count dropped to near zero. His hair fell out. He lost the ability to eat, then to speak. He died on September 15, 1945, fully conscious and aware of what was happening to him [1].
Private Hemmerly, sitting 12 feet away, received a smaller but still significant dose. He developed cancer and died in 1978 at age 62 [3].
The Response
Los Alamos conducted an internal investigation. The conclusion was that Daghlian had violated safety protocols by working alone and by performing the experiment by hand rather than remotely [1]. New rules were issued requiring two scientists present for all criticality work and mandating that experiments stop at an earlier margin from criticality.
But the Demon Core itself — undamaged and still perfectly spherical — remained in use. Criticality experiments continued with the same basic technique. The culture at Los Alamos was one of extraordinary intellectual confidence and institutional bravado. These were the men who had built the atomic bomb in under three years. They trusted their brilliance. The idea that a physics experiment could outrun a physicist's reflexes had not yet been internalized [4].
Daghlian's death was mourned but treated as an avoidable individual error, not a systemic flaw. The same experiments continued on the same core.
Louis Slotin: May 21, 1946
Nine months later, on May 21, 1946, Canadian physicist Louis Slotin was demonstrating a criticality experiment to a group of seven observers in the same Omega Site laboratory. Slotin was 35 years old, a veteran of the Manhattan Project who had personally assembled the core for the Trinity test — the world's first nuclear detonation [2].
Slotin's technique was different from Daghlian's but the principle was identical: bring the core as close to criticality as possible without crossing the threshold. His method involved lowering a beryllium hemisphere — a neutron reflector — over the plutonium core, using a flat-blade screwdriver wedged between the two halves as a spacer to prevent them from closing completely [1]. As he tilted the screwdriver, the gap between the hemispheres narrowed. Neutron counters climbed. The assembly approached criticality.
This procedure was called "tickling the dragon's tail" — a phrase attributed to Richard Feynman [3]. It had no formal safety authorization. It relied entirely on the operator's manual dexterity to maintain a gap of millimeters between safety and catastrophe. Enrico Fermi had warned Slotin months earlier: "You will be dead within a year if you keep doing that" [2].
At 3:20 p.m., the screwdriver slipped. The beryllium hemisphere dropped fully onto the core, closing the assembly. The room was bathed in a blue flash and a pulse of heat. Slotin felt a sour taste in his mouth and a burning sensation in his left hand [1].
He instinctively flipped the hemisphere off the core with his bare hand, stopping the reaction. But the damage was already done.
Slotin's Death
Slotin received approximately 1,000 rem to his upper body — more than double the lethal dose. He was alert enough in the immediate aftermath to tell his colleagues exactly where each of them had been standing, allowing physicists to calculate their individual exposure doses with precision [4]. He then vomited repeatedly, was driven to the hospital, and told the doctors: "I'm afraid I have had it."
Louis Slotin died on May 30, 1946, nine days after the accident. Like Daghlian, he experienced progressive acute radiation syndrome: nausea, then skin blistering, then intestinal lining failure, then bone marrow collapse [3]. His left hand — the one that had flipped the hemisphere — received the highest dose and showed the most severe tissue destruction.
Of the seven observers in the room, all received radiation doses ranging from 35 to 166 rem. The nearest, physicist Alvin Graves, was hospitalized for weeks and suffered permanent vision damage and chronic health problems. At least two others developed radiation-related cancers in subsequent decades [4].
The Name
The plutonium core that killed both Daghlian and Slotin became known as the "Demon Core" among Los Alamos scientists — a nickname that captured the eerie sense that this particular object was malevolent [2]. It was, of course, an ordinary sphere of weapons-grade plutonium. But it had killed two men in two near-identical accidents in nine months, and superstition is a natural human response to repeated lethal coincidence.
The name stuck because it expressed something real about working with nuclear material. These objects were small enough to hold in one hand — the Demon Core weighed less than a bowling ball — yet contained enough energy to level a city. The dissonance between their mundane appearance and apocalyptic potential unsettled even the most hardened weapons scientists [3].
The End of the Demon Core
After Slotin's death, the core was scheduled for destruction in the Operation Crossroads nuclear tests at Bikini Atoll. It was melted down, recast into a slightly different weapon configuration, and shipped to the Marshall Islands in the Pacific [1].
On July 1, 1946 — just six weeks after killing Slotin — the Demon Core was detonated 500 feet above a fleet of 95 target ships in Bikini Lagoon as the "Able" shot. The explosion yielded 23 kilotons, sank five ships, and heavily damaged fourteen others [4]. The Demon Core ended the way it was always intended to — in a nuclear fireball. But it had claimed two lives first.
Legacy
After Slotin's death, Los Alamos permanently banned all hand-assembly criticality experiments. Future tests were conducted remotely, with operators separated from assemblies by concrete shielding, operating mechanical manipulators from behind protective barriers [4]. The era of "tickling the dragon's tail" by hand was over.
The deaths of Daghlian and Slotin became the foundational case studies in nuclear criticality safety, taught in every nuclear engineering program worldwide. They demonstrated that brilliant scientists operating without adequate mechanical safeguards will inevitably make fatal errors — not because they are careless, but because the margin for error is smaller than human motor control can reliably maintain [3].
The lesson is brutally simple: never rely on human reflexes as the last line of defense against a process that moves at the speed of light [1].
Sources
- Los Alamos National Laboratory — *A Review of Criticality Accidents* (LA-13638). Link
- Atomic Heritage Foundation — *The Demon Core*. Link
- Department of Energy — *Criticality Accidents at Los Alamos: Daghlian and Slotin*. Link
- National Nuclear Security Administration — *A History of Criticality Safety*. Link