Criticality accident


A criticality accident is an uncontrolled nuclear fission chain reaction. It is sometimes referred to as a critical excursion, critical power excursion, or divergent chain reaction. Any such event involves the unintended accumulation or arrangement of a critical mass of fissile material, for example enriched uranium or plutonium. Criticality accidents can release potentially fatal radiation doses, if they occur in an unprotected environment.
Under normal circumstances, a critical or supercritical fission reaction should only occur inside a safely shielded location, such as a reactor core or a suitable test environment. A criticality accident occurs if the same reaction is achieved unintentionally, for example in an unsafe environment or during reactor maintenance.
Though dangerous and frequently lethal to humans within the immediate area, the critical mass formed would not be capable of producing a massive nuclear explosion of the type that fission bombs are designed to produce. This is because all the design features needed to make a nuclear warhead cannot arise by chance. In some cases, the heat released by the chain reaction will cause the fissile materials to expand. In such cases, the chain reaction can either settle into a low power steady state or may even become either temporarily or permanently shut down.
In the history of atomic power development, at least 60 criticality accidents have occurred, including 22 in process environments, outside nuclear reactor cores or experimental assemblies, and 38 in small experimental reactors and other test assemblies. Although process accidents occurring outside reactors are characterized by large releases of radiation, the releases are localized. Nonetheless, fatal radiation exposures have occurred to persons close to these events, resulting in 14 fatalities. In a few reactor and critical experiment assembly accidents, the energy released has caused significant mechanical damage or steam explosions.

Physical basis

occurs when sufficient fissile material accumulates in a small volume such that each fission, on average, produces a neutron that in turn strikes another fissile atom causing another fission; this causes the chain reaction to become self-sustaining within the mass of material. In other words, the number of neutrons emitted, over time, exceeds the number of neutrons captured by another nucleus or lost to the environment, resulting in a cascade of increasing nuclear fissions.
Criticality can be achieved by using metallic uranium or plutonium, liquid solutions, or powder slurries. The chain reaction is influenced by range of parameters noted by the acronyms MAGIC MERV and MERMAIDS. Temperature is also a factor.
Calculations can be performed to determine the conditions needed for a critical state, mass, geometry, concentration etc. Where fissile materials are handled in civil and military installations, specially trained personnel are employed to carry out such calculations, and to ensure that all reasonably practicable measures are used to prevent criticality accidents, during both planned normal operations and any potential process upset conditions that cannot be dismissed on the basis of negligible likelihoods.
The assembly of a critical mass establishes a nuclear chain reaction, resulting in an exponential rate of change in the neutron population over space and time leading to an increase in neutron flux. This increased flux and attendant fission rate produces radiation that contains both a neutron and gamma ray component and is extremely dangerous to any unprotected nearby life-form. The rate of change of neutron population depends on the neutron generation time, which is characteristic of the neutron population, the state of "criticality", and the fissile medium.
A nuclear fission creates approximately 2.5 neutrons per fission event on average. Hence, to maintain a stable, exactly critical chain reaction, 1.5 neutrons per fission event must either leak from the system or be absorbed without causing further fissions.
For every 1,000 neutrons released by fission, a small number, typically no more than about 7, are delayed neutrons which are emitted from the fission product precursors, called delayed neutron emitters. This delayed neutron fraction, on the order of 0.007 for uranium, is crucial for the control of the neutron chain reaction in reactors. It is called one dollar of reactivity. The lifetime of delayed neutrons ranges from fractions of seconds to almost 100 seconds after fission. The neutrons are usually classified in 6 delayed neutron groups. The average neutron lifetime considering delayed neutrons is approximately 0.1 sec, which makes the chain reaction relatively easy to control over time. The remaining 993 prompt neutrons are released very quickly, approximately 1 μs after the fission event.
In steady-state operation, nuclear reactors operate at exact criticality. When at least one dollar of reactivity is added above the exact critical point then the chain reaction does not rely on delayed neutrons. In such cases, the neutron population can rapidly increase exponentially, with a very small time constant, known as the prompt neutron lifetime. Thus there is a very large increase in neutron population over a very short time frame. Since each fission event contributes approximately 200 MeV per fission, this results in a very large energy burst as a "prompt-critical spike". This spike can be easily detected by radiation dosimetry instrumentation and "criticality accident alarm system" detectors that are properly deployed.

Accident types

Criticality accidents are divided into one of two categories:
Excursion types can be classified into four categories depicting the nature of the evolution over time:
  1. Prompt criticality excursion
  2. Transient criticality excursion
  3. Exponential excursion
  4. Steady-state excursion
The prompt-critical excursion is characterized by a power history with an initial prompt-critical spike as previously noted, which either self-terminates or continues with a tail region that decreases over an extended period of time. The transient critical excursion is characterized by a continuing or repeating spike pattern after the initial prompt-critical excursion. The longest of the 22 process accidents occurred at Hanford Works in 1962 and lasted for 37.5 hours. The 1999 Tokaimura nuclear accident remained critical for about 20 hours, until it was shut down by active intervention. The exponential excursion is characterized by a reactivity of less than one dollar added, where the neutron population rises as an exponential over time, until either feedback effects or intervention reduce the reactivity. The exponential excursion can reach a peak power level, then decrease over time, or reach a steady-state power level, where the critical state is exactly achieved for a "steady-state" excursion.
The steady-state excursion is also a state which the heat generated by fission is balanced by the heat losses to the ambient environment. This excursion has been characterized by the Oklo natural reactor that was naturally produced within uranium deposits in Gabon, Africa about 1.7 billion years ago.

Recorded incidents

At least sixty criticality accidents have been recorded since 1945. These have caused at least twenty-one deaths: seven in the United States, ten in the Soviet Union, two in Japan, one in Argentina, and one in Yugoslavia. Nine have been due to process accidents, and the others from research reactor accidents.
Criticality accidents have occurred in the context of production and testing of fissile material for both nuclear weapons and nuclear reactors.
DateLocationDescriptionInjuriesFatalitiesRefs
Los AlamosOtto Frisch received a larger than intended dose of radiation when leaning over the original Lady Godiva device for a couple of seconds. He noticed that the red lamps were 'glowing continuously'. Frisch's body had reflected some neutrons back to the device, increasing its neutron multiplication, and it was only by quickly leaning back and away from the device and removing a couple of the uranium blocks that Frisch escaped harm. Afterward he said, "If I had hesitated for another two seconds before removing the material... the dose would have been fatal". On 3 February 1954 and 12 February 1957, accidental criticality excursions occurred causing damage to the device, but fortunately only insignificant exposures to personnel. This original Godiva device was irreparable after the second accident and was replaced by the Godiva II.00
Los AlamosScientist John Bistline was conducting an experiment to determine the effect of surrounding a sub-critical mass of enriched uranium with a water reflector. The experiment unexpectedly became critical when water leaked into the polyethylene box holding the metal. When that happened, the water began to function as a highly effective moderator rather than just a neutron reflector. Three people received non-fatal doses of radiation.30
Los AlamosScientist Harry Daghlian suffered fatal radiation poisoning and died 25 days later after accidentally dropping a tungsten carbide brick onto a sphere of plutonium, which was later nicknamed the demon core. The brick acted as a neutron reflector, bringing the mass to criticality. This was the first known criticality accident causing a fatality.01
Los AlamosScientist Louis Slotin accidentally irradiated himself during a similar incident using the same "demon core" sphere of plutonium responsible for the Daghlian accident. Slotin surrounded the plutonium sphere with two 9-inch diameter hemispherical cups of the neutron-reflecting material beryllium; one above and one below. He was using a screwdriver to keep the cups slightly apart, which kept the assembly subcritical. When the screwdriver accidentally slipped, the cups closed completely around the plutonium, sending the assembly supercritical. Slotin quickly disassembled the device, an act which probably saved the lives of seven fellow scientists nearby. Slotin succumbed to radiation poisoning nine days later. The demon core was melted down and reused in other bomb tests in subsequent years.81
Idaho National LaboratoryThe prototype nuclear jet propulsion reactor HTRE-3 suffered a "power excursion"—an accident that resulted in a partial meltdown, damaging all of the fuel rods in the reactor. This happened during what was supposed to be a low-power run to observe heating rates of reactor components, with the only cooling provided to the reactor coming from a pair of electric blowers. The accident was attributed to misconfigured sensors, not the design. Those sensors gave an incorrect power reading, causing the control rods to be extracted too far. No injuries were reported.00
Oak Ridge, Tennessee Y-12 incidentThe first recorded uranium-processing–related criticality occurred at the Y-12 Plant. During a routine leak test a fissile solution was unknowingly allowed to collect in a 55-gallon drum. The excursion lasted for approximately 20 minutes and resulted in eight workers receiving significant exposure. There were no fatalities, though five were hospitalized for forty-four days. All eight workers eventually returned to work.80
Vinča Nuclear InstituteA criticality excursion in the heavy water RB reactor at the Vinca Nuclear Institute in Vinča, Yugoslavia, killing one person and injuring five. The initial survivors received the first bone marrow transplant in Europe.51
Los AlamosCecil Kelley, a chemical operator working on plutonium purification, switched on a stirrer on a large mixing tank, which created a vortex in the tank. The plutonium, dissolved in an organic solvent, flowed into the center of the vortex. Due to a procedural error, the mixture contained 3.27 kg of plutonium, which reached criticality for about 200 microseconds. Kelley received 3,900 to 4,900 rad according to later estimates. The other operators reported seeing a flash of light and found Kelley outside, saying "I'm burning up! I'm burning up!" He died 35 hours later.01
SL-1, west of Idaho FallsSL-1, a United States Army experimental nuclear power reactor underwent a steam explosion and core disassembly due to improper withdrawal of the central control rod, killing its three operators.03
Wood River JunctionThe facility in Richmond, Rhode Island was designed to recover uranium from scrap material left over from fuel element production. Technician Robert Peabody, intending to add trichloroethene to a tank containing uranium-235 and sodium carbonate to remove organics, added uranium solution instead, producing a criticality excursion. The operator was exposed to a fatal radiation dose of 10,000 rad. Ninety minutes later a second excursion happened when a plant manager returned to the building and turned off the agitator, exposing himself and another administrator to doses of up to 100 rad without ill effect. The operator involved in the initial exposure died 49 hours after the incident.01
MayakThe nuclear fuel processing center in central Russia was experimenting with plutonium purification techniques using different solvents for solvent extraction. Some of these solvents carried over to a tank not intended to hold them, and exceeded the fissile safe limit for that tank. Against procedure a shift supervisor ordered two operators to lower the tank inventory and remove the solvent to another vessel. Two operators were using an "unfavorable geometry vessel in an improvised and unapproved operation as a temporary vessel for storing plutonium organic solution"; in other words, the operators were decanting plutonium solutions into the wrong type—more importantly, shape—of container. After most of the solvent solution had been poured out, there was a flash of light and heat. "Startled, the operator dropped the bottle, ran down the stairs, and from the room." After the complex had been evacuated, the shift supervisor and radiation control supervisor re-entered the building. The shift supervisor then deceived the radiation control supervisor and entered the room of the incident; this was followed by the third and largest criticality excursion that irradiated the shift supervisor with a fatal dose of radiation, possibly due to an attempt by the supervisor to pour the solution down a floor drain.11
:es:Centro Atómico Constituyentes|Centro Atomico ConstituyentesAn operator at the :es:Accidente nuclear del reactor RA-2|RA-2 research reactor in Buenos Aires, Argentina received a fatal radiation dose of 3700 rad while changing the fuel rod configuration with moderating water in the reactor. Two others were injured.21
SarovRussian Federal Nuclear Center senior researcher Alexandr Zakharov received a fatal dose of 4850 rem in a criticality accident.01
TōkaiAt the Japanese uranium reprocessing facility in Ibaraki Prefecture, workers put a mixture of uranyl nitrate solution into a precipitation tank which was not designed to dissolve this type of solution and caused an eventual critical mass to be formed, and resulted in the death of two workers from severe radiation exposure.12

There was speculation although not confirmed within criticality accident experts, that Fukushima 3 suffered a criticality accident. Based on incomplete information about the 2011 Fukushima I nuclear accidents, Dr. Ferenc Dalnoki-Veress speculates that transient criticalities may have occurred there. Noting that limited, uncontrolled chain reactions might occur at Fukushima I, a spokesman for the International Atomic Energy Agency "emphasized that the nuclear reactors won't explode." By 23 March 2011, neutron beams had already been observed 13 times at the crippled Fukushima nuclear power plant. While a criticality accident was not believed to account for these beams, the beams could indicate nuclear fission is occurring. On 15 April, TEPCO reported that nuclear fuel had melted and fallen to the lower containment sections of three of the Fukushima I reactors, including reactor three. The melted material was not expected to breach one of the lower containers, which could cause a massive radioactivity release. Instead, the melted fuel is thought to have dispersed uniformly across the lower portions of the containers of reactors No. 1, No. 2 and No. 3, making the resumption of the fission process, known as a "recriticality", most unlikely.

Observed effects

Blue glow

It has been observed that many criticality accidents emit a blue flash of light.
The blue glow of a criticality accident results from the fluorescence of the excited ions, atoms and molecules of the surrounding medium falling back to unexcited states. This is also the reason electrical sparks in air, including lightning, appear electric blue. The smell of ozone was said to be a sign of high ambient radioactivity by Chernobyl liquidators.
This blue flash or "blue glow" can also be attributed to Cherenkov radiation, if either water is involved in the critical system or when the blue flash is experienced by the human eye. Additionally, if ionizing radiation directly transects the vitreous humor of the eye, Cherenkov radiation can be generated and perceived as a visual blue glow/spark sensation.
It is a coincidence that the color of Cherenkov light and light emitted by ionized air are a very similar blue; their methods of production are different. Cherenkov radiation does occur in air for high-energy particles but not for the lower energy charged particles emitted from nuclear decay.
In a nuclear setting, Cherenkov radiation is instead seen in dense media such as water or in a solution such as uranyl nitrate in a reprocessing plant. Cherenkov radiation could also be responsible for the "blue flash" experienced in an excursion due to the intersection of particles with the vitreous humour within the eyeballs of those in the presence of the criticality. This would also explain the absence of any record of blue light in video surveillance of the more recent incidents.

Heat effects

Some people reported feeling a "heat wave" during a criticality event. It is not known whether this may be a psychosomatic reaction to the realization of what has just occurred, or if it is a physical effect of heating due to radiation emitted by the criticality event.
A review of all of the criticality accidents with eyewitness accounts indicates that the heat waves were only observed when the fluorescent blue glow was also observed. This would suggest a possible relationship between the two, and indeed, one can be potentially identified. In dense air, over 30% of the emission lines from nitrogen and oxygen are in the ultraviolet range, and about 45% are in the infrared range. Only about 25% are in the visible range. Since the skin feels light through its heating of the skin surface, it is possible that this phenomenon can explain the heat wave perceptions. However, this explanation has not been confirmed and may be inconsistent with the intensity of light reported by witnesses compared to the intensity of heat perceived. Further research is hindered by the small amount of data available from the few instances where humans have witnessed these incidents and survived long enough to provide a detailed account of their experiences and observations.

Related terms and concepts