From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
SL-1 Nuclear Meltdown
November 29, 1961: The SL-1 reactor vessel being removed from the reactor building, which acted substantially like the containment building used in modern nuclear facilities. The 60-ton Manitowoc Model 3900 crane had a 5.25-inch (13.3 cm) steel shield with a 9-inch (23 cm) thick lead glass window to protect the operator.
Date3 January 1961
LocationNational Reactor Testing Station, Idaho Falls, Idaho (now Idaho National Laboratory)
Coordinates43°31′06″N 112°49′25″W / 43.5182°N 112.8237°W / 43.5182; -112.8237Coordinates: 43°31′06″N 112°49′25″W / 43.5182°N 112.8237°W / 43.5182; -112.8237
OutcomeINES Level 4 (accident with local consequences)
SL-1 is located in USA West
Location in the western United States
SL-1 is located in Idaho
Location in Idaho, west of Idaho Falls

The SL-1, or Stationary Low-Power Reactor Number One, was a United States Army experimental nuclear power reactor in the United States that underwent a steam explosion and meltdown on January 3, 1961, killing its three operators.[1][2][3][4] The direct cause was the improper withdrawal of the central control rod, responsible for absorbing neutrons in the reactor core. The event is the only reactor accident in U.S. history to have resulted in immediate fatalities.[5] The accident released about 80 curies (3.0 TBq) of iodine-131,[6] which was not considered significant due to its location in the remote high desert of eastern Idaho. About 1,100 curies (41 TBq) of fission products were released into the atmosphere.[7]

The facility, located at the National Reactor Testing Station (NRTS) approximately forty miles (65 km) west of Idaho Falls, Idaho, was part of the Army Nuclear Power Program and was known as the Argonne Low Power Reactor (ALPR) during its design and build phase. It was intended to provide electrical power and heat for small, remote military facilities, such as radar sites near the Arctic Circle, and those in the DEW Line.[8] The design power was 3 MW (thermal),[9] but some 4.7 MW tests were performed in the months prior to the accident. Operating power was 200 kW electrical and 400 kW thermal for space heating.[9]

During the accident the core power level reached nearly 20 GW in just four milliseconds, precipitating the steam explosion.[10][11][12][13]

Design and operations[edit]

From 1954 to 1955, the U.S. Army evaluated their need for nuclear reactor plants that would be operable in remote regions of the Arctic. The reactors were to replace diesel generators and boilers that provided electricity and space heating for the Army's radar stations. The Army Reactors Branch formed the guidelines for the project and contracted with Argonne National Laboratory (ANL) to design, build, and test a prototype reactor plant to be called the Argonne Low Power Reactor (ALPR).[14]

Some of the more important criteria included:

  • All components able to be transported by air[9]
  • All components limited to packages measuring 7.5 by 9 by 20 feet (2.3 m × 2.7 m × 6.1 m) and weighing 20,000 pounds (9,100 kg)[9]
  • Use of standard components
  • Minimal on-site construction[9]
  • Simplicity and reliability[9]
  • Adaptable to the Arctic "permafrost region"[9]
  • 3-year fuel operating lifetime per core loading[14][9]

The prototype was constructed at the NRTS site from July 1957 to July 1958. It went critical on August 11, 1958,[14] became operational on October 24, and was formally dedicated on December 2, 1958.[14] The 3 MW (thermal) boiling water reactor (BWR) used 93.20% highly enriched uranium fuel.[15] It operated with natural circulation, using light water as a coolant (vs. heavy water) and moderator. ANL used its experience from the BORAX experiments to design the BWR. The circulating water system operated at 300 pounds per square inch (2,100 kPa) flowing through fuel plates of uranium-aluminum alloy. The plant was turned over to the U.S. Army for training and operating experience in December 1958 after extensive testing, with Combustion Engineering Incorporated (CEI) acting as the lead contractor beginning February 5, 1959.[16]

CEI was responsible for the actual operation of the SL-1 reactor, for the routine training of military personnel and for developmental research programs.

The Contractor provided at the site a Project Manager, Operations Supervisor, a Test Supervisor, and a technical staff of approximately six personnel. In recent months, the Project Manager spent approximately half time at the site and half time at the contractor's office in Connecticut. In his absence, either the Operations Supervisor or the Test Supervisor was assigned as the Project Manager.

... It was understood, as indicated by testimony before the Board, that CEI would provide supervision on any shifts when non-routine work was carried out.

... the AEC's Idaho Office and the Army Reactors Office clearly believed that the addition of night supervisors when only routine work was involved would defeat a part of the purpose of operating the reactor under the existing arrangement, i.e., to obtain plant operating experience with only military personnel.

— Report on the SL-1 Incident, January 3, 1961, pp. 6 - 7[17]

Trainees in the Army Reactor Training Program included members of the U.S. Army, called cadre, who were the primarily plant operators, although many "Maritime" civilians (see NS Savannah) trained, along with a few U.S. Air Force and U.S. Navy personnel.[16] While plant operation was generally done by the cadre in two-man crews, any development of the reactor was to be supervised directly by CEI staff. CEI decided to perform development work on the reactor as recent as the latter half of 1960 in which the reactor was to be operated at 4.7 MWthermal for a "PL-1 condenser test."[18] As the reactor core aged and boron "poison" strips corroded and flaked off, CEI calculated that about 18% of the boron in the core had been "lost." This resulted in the addition of "cadmium sheets" (also a "poison") on November 11, 1960, two months before the accident. Corrective actions to the flaking off and/or corrosion loss of boron poison strips included a recent[clarification needed] change to the reactor core in which new cadmium sheets were installed "to several tee slot positions to increase reactor shutdown margin."[19] This modification was done on November 15, 1960, some weeks before the accident.

The ALPR before the accident. The large cylindrical building holds the nuclear reactor embedded in gravel at the bottom, the main operating area or operating floor in the middle, and the condenser fan room near the top. Miscellaneous support and administration buildings surround it.

The majority of the plant equipment was located in a cylindrical steel reactor building 38.5 feet (11.7 m) in diameter with an overall height of 48 feet (15 m).[9] The reactor building known as ARA-602 was made of plate steel, most of which had a thickness of 1/4 inch (6 mm). Access to the building was provided by an ordinary door through an enclosed exterior stairwell from ARA-603, the Support Facilities Building. An emergency exit door was also included, with an exterior stairwell going to the ground level.[9] The reactor building was not a pressure-type containment shell as would have been used for reactors located in populated areas. Nevertheless, the building was able to contain most of the radioactive particles released by the eventual explosion.

The reactor core structure was built for a capacity of 59 fuel assemblies, one startup neutron source assembly, and 9 control rods. The core in use, however, had 40 fuel elements and was controlled by 5 cruciform rods.[9] The 5 active rods were in the shape of a plus symbol (+) in cross section: 1 in the center (Rod Number 9), and 4 on the periphery of the active core (Rods 1, 3, 5, and 7).[9] The control rods were made of 60 mils (1.5 mm) thick cadmium, clad with 80 mils (2.0 mm) of aluminum. They had an overall span of 14 inches (36 cm) and an effective length of 32 inches (81 cm).[9] The 40 fuel assemblies were composed of 9 fuel plates each.[9] The plates were 120 mils (3.0 mm) thick consisting of 50 mils (1.3 mm) of uranium-aluminum alloy "meat" covered by 35 mils (0.89 mm) of X-8001 aluminum cladding.[9] The meat was 25.8 inches (66 cm) long and 3.5 inches (8.9 cm) wide. The water gap between fuel plates was 310 mils (7.9 mm).[9] Water channels within the control rod shrouds was 0.5 inches (13 mm). The initial loading of the 40 assembly core was highly enriched with 93.2% uranium-235 and contained 31 pounds (14 kg) of U-235.[9]

The deliberate choice of a smaller fuel loading made the region near the center more active than it would have been with 59 fuel assemblies. The four outer control rods were not even used in the smaller core after tests concluded they were not necessary.[9][17] In the operating SL-1 core, Rods Number 2, 4, 6, and 8 were dummy rods, had cadmium shims after November 11, 1960, or were filled with test sensors, and were shaped like the capital letter T.[18] The effort to minimize the size of the core gave the central rod an abnormally large reactivity worth.

Accident and response[edit]

On December 21, 1960, the reactor was shut down for maintenance, calibration of the instruments, installation of auxiliary instruments, and installation of 44 flux wires to monitor the neutron flux levels in the reactor core. The wires were made of aluminum, and contained slugs of aluminum–cobalt alloy.

On January 3, 1961, the reactor was being prepared for restart after a shutdown of eleven days over the holidays. Maintenance procedures required that the main central control rod be manually withdrawn a few inches to reconnect it to its drive mechanism. At 9:01 p.m., this rod was suddenly withdrawn too far, causing SL-1 to go prompt critical instantly. In four milliseconds, the heat generated by the resulting enormous power excursion caused fuel inside the core to melt and to explosively vaporize. The expanding fuel produced an extreme pressure wave that blasted water upward, striking the top of the reactor vessel with a peak pressure of 10,000 pounds per square inch (69,000 kPa). The slug of water was propelled at about 159 feet per second (48 m/s) with average pressure of around 500 pounds per square inch (3,400 kPa).[15] This extreme form of water hammer propelled the entire reactor vessel upward at about 27 feet per second (8.2 m/s), while the shield plugs were ejected at about 85 feet per second (26 m/s).[15] With 6 holes on the top of the reactor vessel, high pressure water and steam sprayed the entire room with radioactive debris from the damaged core. A later investigation concluded that the 26,000-pound (12,000 kg) (or 13 short tons) vessel had jumped 9 feet 1 inch (2.77 m), parts of which struck the ceiling of the reactor building before settling back into its original location,[11][20][15] and depositing insulation and gravel on the operating floor.[15] If not for the vessel's #5 seal housing hitting the overhead crane, the pressure vessel had enough upward momentum to rise about 10 feet (3.0 m).[15] The entire time for the excursion, steam explosion, and vessel movement took between two and four seconds.[15]

The spray of water and steam knocked two operators onto the floor, killing one and severely injuring another. The No. 7 shield plug from the top of the reactor vessel impaled the third man through his groin and exited his shoulder, pinning him to the ceiling.[11] The victims were Army Specialists John A. Byrnes (age 22) and Richard Leroy McKinley (age 27), and Navy Seabee Construction Electrician First Class (CE1) Richard C. Legg (age 26).[21][22][23] It was later established by author Todd Tucker that Byrnes (the reactor operator) had lifted the rod and caused the excursion; Legg (the shift supervisor) was standing on top of the reactor vessel and was impaled and pinned to the ceiling; and, McKinley (the trainee) stood nearby. Only McKinley was found alive, though unconscious and in deep shock, by rescuers.[11] This was consistent with the analysis of the SL-1 Board of Investigation[24] and consistent with the results of the autopsy, which suggested that Byrnes and Legg died instantly, while McKinley showed signs of diffuse bleeding within his scalp, indicating he survived approximately two hours before succumbing to his wounds.[25] All three men succumbed to injuries from physical trauma;[11][25] however, the radiation from the nuclear excursion would have given the men no chance of survival even had they not been killed by the explosion stemming from the criticality accident.

Reactor principles and events[edit]

Several "kinetic" factors affect the rate at which the power (heat) produced in a nuclear reactor responds to changes in the position of a control rod. Other features of the design govern how rapidly heat is transferred from the reactor fuel to the coolant.

The nuclear chain reaction has a positive feedback component whenever a critical mass is created; specifically, excess neutrons are produced for every fission. Inside a nuclear reactor, these excess neutrons must be controlled as long as a critical mass exists. The most significant and effective control mechanism is the use of control rods to absorb the excess neutrons. Other controls include the size and shape of the reactor and the presence of neutron reflectors in and around a core. Changing the amount of absorption or reflection of neutrons will affect neutron flux, and therefore, the power of the reactor.

One kinetic factor is the tendency of most light-water-moderated reactor (LWR) designs to have negative moderator temperature and void coefficients of reactivity. (Due to the low density of steam, pockets of water vapor are known as "voids" in a LWR.) A negative reactivity coefficient means that as the water moderator heats up, molecules move farther apart (water expands and eventually boils) and neutrons are less likely to be slowed by collisions to energies favorable for inducing fission in the fuel. Because of these negative feedback mechanisms, most LWRs will naturally tend to decrease their rate of fissioning in response to additional heat produced within the reactor core. If enough heat is produced that water boils inside the core, fissions in that vicinity will drastically decrease.

However, when power output from the nuclear reaction increases rapidly, it may take longer for the water to heat up and boil than it does for the voids to cause the nuclear reactions to decrease. In such an event, reactor power can grow rapidly without any negative feedback from the expansion or boiling of the water, even if it is in a channel just 1 cm (0.39 in) away. Dramatic heating will occur to the nuclear fuel, leading to melting and vaporization of the metals within the core. Rapid expansion, increases in pressure, and failure of core components may lead to the destruction of the nuclear reactor, as was the case with SL-1. As the energy of expansion and heat travel from the nuclear fuel to the water and the vessel, it becomes likely that the nuclear reaction will shut down, either from the lack of sufficient moderator or from the fuel expanding beyond the realm of a critical mass. In the post-accident analysis of SL-1, scientists determined that the two shutdown mechanisms were almost equally matched (see below).

Another relevant kinetics factor is the contribution of what are called delayed neutrons to the chain reaction in the core. Most neutrons (the prompt neutrons) are produced nearly instantaneously via fission. But a few—approximately 0.7 percent in a U-235-fueled reactor operating at steady-state—are produced through the relatively slow radioactive decay of certain fission products. (These fission products are trapped inside the fuel plates in close proximity to the uranium-235 fuel.) Delayed production of a fraction of the neutrons is what enables reactor power changes to be controllable on a time scale that is amenable to humans and machinery.[26]

In the case of an ejected control assembly or poison, it is possible for the reactor to become critical on the prompt neutrons alone (i.e. prompt critical). When the reactor is prompt critical, the time to double the power is on the order of 10 microseconds. The duration necessary for temperature to follow the power level depends on the design of the reactor core. Typically, the coolant temperature lags behind the power by 3 to 5 seconds in a conventional LWR. In the SL-1 design, it was about 6 milliseconds before steam formation started.[15]

SL-1 was constructed with a main central control rod that was capable of producing a very large excess reactivity if it were completely removed. The extra rod worth was in part due to the decision to load only 40 of the 59 fuel assemblies with nuclear fuel, thus making the prototype reactor core more active in the center. In normal operation control rods are withdrawn only enough to cause sufficient reactivity for a sustained nuclear reaction and power generation. In this accident, however, the reactivity addition was sufficient to take the reactor prompt critical within a time estimated at 4 milliseconds.[27] That was too fast for the heat from the fuel to get through the aluminum cladding and boil enough water to fully stop the power growth in all parts of the core via negative moderator temperature and void feedback.[15][27]

Post-accident analysis concluded that the final control method (i.e., dissipation of the prompt critical state and the end of the sustained nuclear reaction) occurred by means of catastrophic core disassembly: destructive melting, vaporization, and consequent conventional explosive expansion of the parts of the reactor core where the greatest amount of heat was being produced most quickly. It was estimated that this core heating and vaporization process happened in about 7.5 milliseconds, before enough steam had been formed to shut down the reaction, beating the steam shutdown by a few milliseconds. A key statistic makes it clear why the core blew apart: the reactor designed for a 3 MW power output operated momentarily at a peak of about 20 GW, a power density over 6,000 times higher than its safe operating limit.[13] This criticality accident is estimated to have produced 4.4 x 1018 fissions,[13] or about 133 megajoules (32 kilograms of TNT).[27]

Events after the power excursion[edit]

Checking for radioactive contamination on nearby Highway 20

There were no other people at the reactor site. The ending of the nuclear reaction was caused solely by the design of the reactor and the basic physics of heated water and core elements melting, separating the core elements and removing the moderator.

Heat sensors above the reactor set off an alarm at the central test site security facility at 9:01 p.m. MST, the time of the accident. False alarms had occurred in the morning and afternoon that same day. The first response crew of six firemen (Ken Dearden Asst Chief, Mel Hess Lt., Bob Archer, Carl Johnson, Egon Lamprecht, Gerald Stuart, & Vern Conlon) arrived nine minutes later, expecting another false alarm.[28] They initially noticed nothing unusual, with only a little steam rising from the building, normal for the cold 6 °F (−14 °C) night. The control building appeared normal. The firefighters entered the reactor building and noticed a radiation warning light. Their radiation detectors jumped sharply to above their maximum range limit as they were climbing the stairs to SL-1's floor level. They peered into the reactor room before withdrawing.[28]

At 9:17 p.m., a health physicist arrived. He and a fireman, both wearing air tanks and masks with positive pressure in the mask to force out any potential contaminants, approached the reactor building stairs. Their detectors read 25 röntgens per hour (R/hr) as they started up the stairs, and they withdrew.

Some minutes later, a health physics response team arrived with radiation meters capable of measuring gamma radiation up to 500 R/hr and full-body protective clothing. One health physicist and two firefighters ascended the stairs and, from the top, saw damage in the reactor room. With the meter showing maximum scale readings, they withdrew rather than approach the reactor more closely and risk further exposure.

The stretcher rig. Army volunteers from a special Chemical Radiological Unit at Dugway Proving Ground practiced before a crane inserted the rig into the SL-1 reactor building to collect the body of the man (Legg) pinned to the ceiling directly above the reactor vessel.

Around 10:30 p.m. MST, the supervisor for the contractor running the site (Combustion Engineering) and the chief health physicist arrived. They entered the reactor building around 10:45 pm and found two mutilated men soaked with water: one clearly dead (Byrnes), the other moving slightly (McKinley) and moaning. With one entry per person and a 1-minute limit, a team of 5 men with stretchers recovered the operator who was still breathing around 10:50; he did not regain consciousness and died of his head injury at about 11 p.m. Even stripped, his body was so contaminated that it was emitting about 500 R/hr. Meanwhile, the third man was discovered about 11:38 p.m., impaled to the ceiling. With all potential survivors now recovered, safety of rescuers took precedence and work was slowed to protect them.

The third man was discovered last because he was pinned to the ceiling above the reactor by a shield plug and not easily recognizable.[11] On the night of January 4, a team of six volunteers used a plan involving teams of two to recover the body of Byrnes. On January 9, in relays of two at a time, a team of ten men, allowed no more than 65 seconds exposure each, used sharp hooks on the end of long poles to pull Legg's body free of the shield plug, dropping it onto a 5-by-20-foot (1.5 by 6.1 m) stretcher attached to a crane outside the building.[11]

Radioactive gold 198Au from the man's gold watch buckle and copper 64Cu from a screw in a cigarette lighter subsequently proved that the reactor had indeed gone prompt critical. Prior to the discovery of neutron-activated elements in the men's belongings, scientists had doubted that a nuclear excursion had occurred, believing the reactor was inherently safe. These findings ruled out early speculations that a chemical explosion caused the accident.[20]

The bodies of all three were buried in lead-lined caskets sealed with concrete and placed in metal vaults with a concrete cover. Some highly radioactive body parts were buried in the Idaho desert as radioactive waste. Army Specialist Richard Leroy McKinley is buried in section 31 of Arlington National Cemetery.

Some sources and eyewitness accounts confuse the names and positions of each victim.[11] In Idaho Falls: The untold story of America's first nuclear accident,[29] the author indicates that initial rescue teams identified Byrnes as the man found initially alive, believing that Legg's body was the one found next to the reactor shield and recovered the night after the accident, and that McKinley was impaled by the control rod to the ceiling directly above the reactor. This misidentification, caused by the severe blast injuries to the victims, was later rectified during the autopsies, but this would cause confusion for some time.[29]


One of the required maintenance procedures called for the central control rod to be manually withdrawn approximately 4 inches (10 cm) in order to attach it to the automated control mechanism from which it had been disconnected. Post-accident calculations estimate that the main control rod was actually withdrawn approximately 26 inches (66 cm), causing the reactor to go prompt critical, which resulted in the steam explosion. The fuel, portions of the fuel plates, and water surrounding the fuel plates vaporized in the extreme heat. The expansion caused by this heating process caused water hammer as water was accelerated upwards toward the reactor vessel head, producing peak pressures of 10,000 pounds per square inch (69,000 kPa) on the head of the reactor vessel when air and then water struck the head at 160 feet per second (50 m/s).[27]

The water hammer not only caused extreme physical damage and distortion of the reactor vessel, it also caused the shield plugs of the vessel to be ejected, one of which impaled Legg. The most surprising and unforeseen evidences of the steam explosion and water hammer were the impressions made on the ceiling above the reactor vessel when it jumped 9 feet 1 inch (2.77 m) in the air before settling back into its prior location. The post-accident analysis also concluded that the reactor vessel was dry, since most of the water and steam had been either ejected immediately or evaporated due to the heat inside the reactor. A borescope was used to confirm this prior to the removal of the reactor.

It was water hammer that caused the physical damage to the reactor, the deaths of personnel who stood atop and nearby, and the release of radioactive isotopes to the environment. One of the lessons learned from SL-1 was that there is an extreme water hammer hazard whenever a shutdown reactor is cooled to room temperature and there is an air gap between the top of the water and the reactor vessel head. One of the recommendations in the analysis of the accident was that shutdown reactors be filled to the top with water so that a power excursion could not induce such a powerful water hammer. Air is not dense enough to appreciably slow water, while water (being nearly incompressible) is able to distribute explosive forces and limit peak pressure. The extra water is also a very effective radiation shield for those who are directly above the vessel. Written procedures at SL-1 had included a directive to pump down the level of water in the reactor prior to the maintenance procedure that destroyed it.

The most common theories proposed for the withdrawal of the rod are (1) sabotage or suicide by one of the operators, (2) a suicide-murder involving an affair with the wife of one of the other operators, (3) inadvertent withdrawal of the main control rod, or (4) an intentional attempt to "exercise" the rod (to make it travel more smoothly within its sheath).[30][31] The maintenance logs do not address what the technicians were attempting to do, and thus the actual cause of the accident will never be known. The investigation took almost two years to complete.

Investigators analyzed the flux wires installed during the maintenance to determine the power output level. They also examined scratches on the central control rod. Using this data, they concluded that the central rod had been withdrawn 26.25 inches (66.7 cm).[20] The reactor would have been critical at 23 inches (58.4 cm), and it took approximately 100 ms for the rod to travel the final 3.25 inches (8.3 cm). Once this was calculated, experiments were conducted with an identically weighted mock control rod to determine whether it was possible or feasible for one or two men to have performed this. Experiments included a simulation of the possibility that the 48-pound (22 kg)[32] rod was stuck and one man freed it himself, reproducing the scenario that investigators considered the best explanation: Byrnes broke the control rod loose and withdrew it accidentally, killing all three men.[11]

At SL-1, control rods would get stuck in the control rod channel sporadically. Numerous procedures were conducted to evaluate control rods to ensure they were operating properly. There were rod drop tests and scram tests of each rod, in addition to periodic rod exercising and rod withdrawals for normal operation. From February 1959 to November 18, 1960, there were 40 cases of a stuck control rod for scram and rod drop tests and about a 2.5% failure rate. From November 18, 1960, to December 23, 1960, there was a dramatic increase in stuck rods, with 23 in that time period and a 13.0% failure rate. Besides these test failures, there were an additional 21 rod-sticking incidents from February 1959 to December 1960; 4 of these had occurred in the last month of operation during routine rod withdrawal. The central control rod, No. 9, had the best operational performance record even though it was operated more frequently than any of the other rods.

Rod sticking has been attributed to misalignment, corrosion product build-up, bearing wear, clutch wear, and drive mechanism seal wear. Many of the failure modes that caused a stuck rod during tests (like bearing and clutch wear) would only apply to a movement performed by the control rod drive mechanism. Since the No. 9 rod is centrally located, its alignment may have been better than Nos. 1, 3, 5, and 7, which were more prone to sticking. After the accident, logbooks and former plant operators were consulted to determine if there had been any rods stuck during the reassembly operation that Byrnes was performing. One person had performed this about 300 times, and another 250 times; neither had ever felt a control rod stick when being manually raised during this procedure. Furthermore, no one had ever reported a stuck rod during manual reconnection.

The mechanical and material evidence, combined with the nuclear and chemical evidence, forced them to believe that the central control rod had been withdrawn very rapidly. ... The scientists questioned the [former operators of SL-1]: "Did you know that the reactor would go critical if the central control rod were removed?" Answer: "Of course! We often talked about what we would do if we were at a radar station and the Russians came. We'd yank it out."

— Susan M. Stacy, Proving the Principle, [20]


The accident caused this design to be abandoned and future reactors to be designed so that a single control rod removal would not have the ability to produce the very large excess reactivity that was possible with this design. Today this is known as the "one stuck rod" criterion and requires complete shutdown capability even with the most reactive rod stuck in the fully withdrawn position. The reduced excess reactivity limits the possible size and speed of the power surge. The "one stuck rod" criterion did not originate as a result of the SL-1 accident. It was, in fact, a hard and fast design criterion long before the SL-1, from the beginning of the Naval Reactors program, under the leadership of Admiral Hyman Rickover. This design criterion started with the USS Nautilus, and continued throughout subsequent submarine and surface ship designs, and with the Shippingport civilian nuclear plant. It continues to be a requirement for all U.S. reactor designs to this day.[contradictory]

Although portions of the center of the reactor core had been vaporized briefly, very little corium was recovered. The fuel plates showed signs of catastrophic destruction leaving voids, but "no appreciable amount of glazed molten material was recovered or observed." Additionally, "There is no evidence of molten material having flowed out between the plates." It is believed that rapid cooling of the core was responsible for the small amount of molten material. There was insufficient heat generated for any corium to reach or penetrate the bottom of the reactor vessel. The reactor vessel was removed on November 29, 1961, without accident. The only holes in the bottom of the vessel were the ones drilled to allow the insertion of borescopes to determine the condition of the melted core.

Even without an engineered containment building like those used today, the SL-1 reactor building contained most of the radioactivity, though iodine-131 levels on plant buildings during several days of monitoring reached fifty times background levels downwind. Radiation surveys of the Support Facilities Building, for example, indicated high contamination in halls, but light contamination in offices.

Radiation exposure limits prior to the accident were 100 röntgens to save a life and 25 to save valuable property. During the response to the accident, 22 people received doses of 3 to 27 Röntgens full-body exposure.[33] Removal of radioactive waste and disposal of the three bodies eventually exposed 790 people to harmful levels of radiation.[34] In March 1962, the Atomic Energy Commission awarded certificates of heroism to 32 participants in the response.

The documentation and procedures required for operating nuclear reactors expanded substantially, becoming far more formal as procedures that had previously taken two pages expanded to hundreds. Radiation meters were changed to allow higher ranges for emergency response activities.

After a pause for evaluation of procedures, the Army continued its use of reactors, operating the Mobile Low-Power Reactor (ML-1), which started full power operation on February 28, 1963, becoming the smallest nuclear power plant on record to do so. This design was eventually abandoned after corrosion problems. While the tests had shown that nuclear power was likely to have lower total costs, the financial pressures of the Vietnam War caused the Army to favor lower initial costs and it stopped the development of its reactor program in 1965, although the existing reactors continued operating (MH-1A until 1977).


General Electric Corporation was tasked with the removal of the reactor vessel and the dismantling and cleanup of the contaminated buildings at the SL-1 project site.[15] The site was cleaned in 1961 to 1962, removing the bulk of the contaminated debris and burying it.[15] The massive cleanup operation included the transport of the reactor vessel to a nearby "hot shop" for extensive analysis.[15] Other items of less importance were either disposed of or transported to decontamination sites for various kinds of cleaning. About 475 people took part in the SL-1 site cleanup, including volunteers from the U.S. Army and the Atomic Energy Commission.[15]

The recovery operation included clearing the operating room floor of radioactive debris. The extremely high radiation areas surrounding the reactor vessel and the fan room directly above it contributed to the difficulty of recovering the reactor vessel. Remotely operated equipment, cranes, boom trucks, and safety precautions had to be developed and tested by the recovery team. Radiation surveys and photographic analysis was used to determine what items needed to be removed from the building first.[15] Powerful vacuum cleaners, operated manually by teams of men, collected vast quantities of debris.[15] The manual overhead crane above the operating floor was used to move numerous heavy objects weighing up to 19,600 pounds (8,900 kg) for them to be dumped out onto the ground outside.[15] Hot spots up to 400 R/hr were discovered and removed from the work area.

With the operating room floor relatively clean and radiation fields manageable, the manual overhead crane was employed to do a trial lift of the reactor vessel.[15] The crane was fitted with a dial-type load indicator and the vessel was lifted a few inches. The successful test found that the estimated 23,000 pounds (10,000 kg) vessel plus an unknown amount of debris weighed about 26,000 pounds (12,000 kg). After removing a large amount of the building structure above the reactor vessel, a 60-ton Manitowoc Model 3900 crane lifted the vessel out of the building into an awaiting transport cask attached to a tractor-trailer combination with a low-boy 60-ton capacity trailer.[15] After raising or removing 45 power lines, phone lines, and guy wires from the proposed roadway, the tractor-trailer, accompanied by numerous observers and supervisors, proceeded at about 10 mph (16 km/h) to the ANP Hot Shop (originally associated with the Aircraft Nuclear Propulsion program), located in a remote area of the NRTS known as Test Area North, about 35 miles (56 km) away.[15]

A burial ground was constructed approximately 1,600 feet (500 m) northeast of the original site of the reactor. It was opened on May 21, 1961.[14] Burial of the waste helped minimize radiation exposure to the public and site workers that would have resulted from transport of contaminated debris from SL-1 to the Radioactive-Waste Management Complex over 16 miles (26 km) of public highway. The original cleanup of the site took about 18 months. The entire reactor building, contaminated materials from nearby buildings, and soil and gravel contaminated during cleanup operations were disposed of in the burial ground. The majority of buried materials consist of soils and gravel.[35][36]

SL-1 burial site in 2003, capped with rip rap

Recovered portions of the reactor core, including the fuel and all other parts of the reactor that were important to the accident investigation, were taken to the ANP Hot Shop for study. After the accident investigation was complete, the reactor fuel was sent to the Idaho Chemical Processing Plant for reprocessing. The reactor core minus the fuel, along with the other components sent to the Hot Shop for study, was eventually disposed of at the Radioactive Waste Management Complex.[35]

The remains of the SL-1 reactor are now buried near the original site at 43°31′17.8″N 112°49′04.8″W / 43.521611°N 112.818000°W / 43.521611; -112.818000.[37] The SL-1 burial ground consists of three excavations, in which a total volume of 99,000 cubic feet (2800 m3) of contaminated material was deposited. The excavations were dug as close to basalt as the equipment used would allow and ranges from 8 to 14 feet (2.4 to 4.3 m) in depth. At least 2 feet (0.6 m) of clean backfill was placed over each excavation. Shallow mounds of soil over the excavations were added at the completion of cleanup activities in September 1962. The site and burial mound are collectively known as United States Environmental Protection Agency Superfund Operable Unit 5-05.[35][38]

Numerous radiation surveys and cleanup of the surface of the burial ground and surrounding area have been performed in the years since the SL-1 accident. Aerial surveys were performed by EG&G Las Vegas in 1974, 1982, 1990, and 1993. The Radiological and Environmental Sciences Laboratory conducted gamma radiation surveys every 3 to 4 years between 1973 and 1987 and every year between 1987 and 1994. Particle-picking at the site was performed in 1985 and 1993. Results from the surveys indicated that cesium-137 and its progeny (decay products) are the primary surface-soil contaminants. During a survey of surface soil in June 1994, "hot spots," areas of higher radioactivity, were found within the burial ground with activities ranging from 0.1 to 50 milliroentgen (mR)/hour. On November 17, 1994, the highest radiation reading measured at 2.5 feet (0.75 m) above the surface at the SL-1 burial ground was 0.5 mR/hour; local background radiation was 0.2 mR/hour. A 1995 assessment by the EPA recommended that a cap be placed over the burial mounds. The primary remedy for SL-1 was to be containment by capping with an engineered barrier constructed primarily of native materials.[35] This remedial action was completed in 2000 and first reviewed by the EPA in 2003.[38]

Movies and books[edit]

Animation of the film produced by the Atomic Energy Commission, available from The Internet Archive.

The U.S. Government produced a film about the accident for internal use in the 1960s. The video was subsequently released and can be viewed at The Internet Archive[39] and YouTube. SL-1 is the title of a 1983 movie, written and directed by Diane Orr and C. Larry Roberts, about the nuclear reactor explosion.[34] Interviews with scientists, archival film, and contemporary footage, as well as slow-motion sequences, are used in the film.[40][41] The events of the accident are also the subject of one book: Idaho Falls: The untold story of America's first nuclear accident (2003)[29] and 2 chapters in Proving the Principle - A History of The Idaho National Engineering and Environmental Laboratory, 1949–1999 (2000).[42]

In 1975, the anti-nuclear book We Almost Lost Detroit, by John G. Fuller was published, referring at one point to the Idaho Falls accident. Prompt Critical is the title of a 2012 short film, viewable on YouTube, written and directed by James Lawrence Sicard, dramatizing the events surrounding the SL-1 accident.[43] A documentary about the accident was shown on the History Channel.[44]

A safety poster designed for engineering offices depicting the melted SL-1 reactor core.[45]

Another author, Todd Tucker, studied the accident and published a book detailing the historical aspects of nuclear reactor programs of the U.S. military branches. Tucker used the Freedom of Information Act to obtain reports, including autopsies of the victims, writing in detail how each person died and how parts of their bodies were severed, analyzed, and buried as radioactive waste.[11] The autopsies were performed by the same pathologist known for his work following the Cecil Kelley criticality accident. Tucker explains the reasoning behind the autopsies and the severing of victims' body parts, one of which gave off 1,500 R/hour on contact. Because the SL-1 accident killed all three of the military operators on site, Tucker calls it "the deadliest nuclear reactor incident in U.S. history."[46]

See also[edit]


  1. ^ "3 die in reactor blast". Spokane Daily Chronicle. (Washington). Associated Press. January 4, 1961. p. 1.
  2. ^ Hale, Steve (January 4, 1961). "3 killed in severe blast at Idaho A-reactor site". Deseret News. (Salt Lake City, Utah). p. A1.
  3. ^ "Reactor blast kills three, pours out radiation". Lewiston Morning Tribune. (Idaho). Associated Press. January 5, 1961. p. 1.
  4. ^ "3 technicians die in reactor blast". Spokesman-Review. (Spokane, Washington). Associated Press. January 5, 1961. p. 2.
  5. ^ Stacy, Susan M. (2000). "Chapter 16: The Aftermath" (PDF). Proving the Principle: A History of The Idaho National Engineering and Environmental Laboratory, 1949-1999. U.S. Department of Energy, Idaho Operations Office. pp. 150–157. ISBN 0-16-059185-6.
  6. ^ The Nuclear Power Deception Table 7: Some Reactor Accidents
  7. ^ Horan, J. R., and J. B. Braun, 1993, Occupational Radiation Exposure History of Idaho Field Office Operations at the INEL, EGG-CS-11143, EG&G Idaho, Inc., October, Idaho Falls, Idaho.
  8. ^ "Idaho: Runaway Reactor". Time. January 13, 1961. Retrieved July 30, 2010.
  9. ^ a b c d e f g h i j k l m n o p q r Design of the Argonne Low Power Reactor (ALPR), ANL-6076 Reactor Technology, Grant, Hamer, Hooker, Jorgensen, Kann, Lipinski, Milak, Rossin, Shaftman, Smaardyk, Treshow, May 1961, University of Chicago, Argonne National Laboratory.
  10. ^ Steve Wander (editor) (February 2007). "Supercritical" (PDF). System Failure Case Studies. NASA. 1 (4). Archived from the original (PDF) on 2007-11-27. Retrieved 2007-10-05.CS1 maint: extra text: authors list (link)
  11. ^ a b c d e f g h i j Tucker, Todd (2009). Atomic America: How a Deadly Explosion and a Feared Admiral Changed the Course of Nuclear History. New York: Free Press. ISBN 978-1-4165-4433-3. See summary: [1]
  12. ^ LA-3611 A Review of Criticality Accidents, William R. Stratton, Los Alamos Scientific Laboratory, 1967
  13. ^ a b c LA-13638 A Review of Criticality Accidents (2000 Revision), Thomas P. McLaughlin, et al., Los Alamos National Laboratory, 2000.
  14. ^ a b c d e SEC-00219, Petition Evaluation Report, Idaho National Laboratory (INL), Revision 2, NIOSH/ORAU, Idaho National Laboratory, March 2017
  15. ^ a b c d e f g h i j k l m n o p q r s IDO-19311 Final Report of SL-1 Recovery Operation, Idaho Test Station, General Electric Corporation, July 27, 1962.
  16. ^ a b IDO-19012, CEND-82, SL-1 Annual Operating Report, Feb. 1959 - Feb 1960, Canfield, Vallario, Crudele, Young, Rausch, Combustion Engineering Nuclear Division, May 1, 1960.
  17. ^ a b Report on the SL-1 Incident, January 3, 1961, The General Manager's Board of Investigation, For Release in Newspapers dated Sunday, Curtis A. Nelson, Clifford Beck, Peter Morris, Donald Walker, Forrest Western, June 11, 1961.
  18. ^ a b Joint Committee on Atomic Energy Congress (1961), SL-1 Accident Atomic Energy Commission Investigation Board Report, Joint Committee on Atomic Energy Congress of the United States, Washington, DC, available at: [2] (accessed September 29, 2018).
  19. ^ IDO-19024 SL-1 Annual Operating Report, February 1960 - January 3, 1961 Combustion Engineering Nuclear Division, CEND-1009, W. B. Allred, June 15, 1961.
  20. ^ a b c d Stacy, Susan M. (2000). Proving the Principle - A History of The Idaho National Engineering and Environmental Laboratory, 1949-1999 (PDF). U.S. Department of Energy, Idaho Operations Office. ISBN 0-16-059185-6. Archived from the original (PDF) on 2011-08-07. Chapter 15.
  21. ^ "Nuclear Experts Probe Fatal Reactor Explosion". Times Daily. January 5, 1961. Retrieved July 30, 2010.
  22. ^ "Richard Legg" (JPEG). Find A Grave. 14 May 2011. Retrieved 5 March 2013.
  23. ^ Spokane Daily Chronicle - Jan 4, 1961. The article notes that Byrnes was a "Spec. 5" from Utica, New York, McKinley was a "Spec. 4" from Kenton, Ohio, Legg was a "Navy electrician L.C." from Roscommon, Michigan.
  24. ^ Final Report of SL-1 Accident Investigation Board, SL-1 Board of Investigation, Curtis A. Nelson, Atomic Energy Commission, Joint Committee on Atomic Energy, September 5, 1962 (See Annual Report to Congress - U.S. Atomic Energy Commission, 1962, Appendix 8, pp. 518 to 523)
  25. ^ a b LAMS-2550 SL-1 Reactor Accident Autopsy Procedures and Results, Clarence Lushbaugh, et al., Los Alamos Scientific Laboratory, June 21, 1961.
  26. ^ Lamarsh, John R.; Baratta, Anthony J. (2001). Introduction to Nuclear Engineering. Upper Saddle River, New Jersey: Prentice Hall. p. 783. ISBN 0-201-82498-1.
  27. ^ a b c d IDO-19313: Additional Analysis of the SL-1 Excursion Archived 2011-09-27 at the Wayback Machine Final Report of Progress July through October 1962, November 21, 1962, Flight Propulsion Laboratory Department, General Electric Company, Idaho Falls, Idaho, U.S. Atomic Energy Commission, Division of Technical Information.
  28. ^ a b Berg, Sven (December 12, 2009). "Nuclear accident still mystery to rescue worker". The Argus Observer. Retrieved April 6, 2015.
  29. ^ a b c McKeown, William (2003). Idaho Falls: The Untold Story of America's First Nuclear Accident. Toronto: ECW Press. ISBN 978-1-55022-562-4., [3]
  30. ^ ATOMIC CITY, by Justin Nobel Archived 2012-05-22 at the Wayback Machine Tin House Magazine, Issue #51, Spring, 2012.
  31. ^ A Nuclear Family, By Maud Newton The New York Times Magazine, April 1, 2012.
  32. ^ Grant, N. R.; Hamer, E. E.; Hooker, H. H.; Jorgensen, G. L.; Kann, W. J.; Lipinski, W. C.; Milak, G. C.; Rossin, A. D.; Shaftman, D. H. (1961-05-01). "Design of the Argonne Low Power Reactor (Alpr)": 75. doi:10.2172/4014868. OSTI 4014868. Cite journal requires |journal= (help)
  33. ^ Johnston, Wm. Robert. "SL-1 reactor excursion, 1961". Johnston's Archive. Retrieved 30 July 2010.
  34. ^ a b Maslin, Janet (March 21, 1984). "Sl-1 (1983): Looking at Perils of Toxicity". The New York Times. Retrieved July 30, 2010.
  35. ^ a b c d EPA,OSWER,OSRTI, US. "Superfund - US EPA" (PDF). US EPA.CS1 maint: multiple names: authors list (link)
  36. ^ Record of Decision, Stationary Low-Power Reactor-1 and Boiling Water Reactor Experiment-I Burial Grounds (Operable Units 5-05 and 6-01), and 10 No Action Sites (Operable Units 5-01, 5-03, 5-04, and 5-11), January 1996.
  37. ^ "2003 Annual Inspection Summary for the Stationary Low-Power Reactor-1 Burial Ground" (PDF). INEEL.
  38. ^ a b 2003 Annual Inspection Summary for the Stationary Low-Power Reactor Burial Ground, Operable Unit 5-05
  39. ^ "SL-1 The Accident: Phases I and II".
  40. ^ SL-1 on IMDb
  41. ^ "Movie Reviews".
  42. ^ Stacy, Susan M. (2000). Proving the Principle: A History of The Idaho National Engineering and Environmental Laboratory, 1949-1999. U.S. Department of Energy, Idaho Operations Office. ISBN 0-16-059185-6.
  43. ^ Prompt Critical on YouTube by James Lawrence Sicard.
  44. ^ SL-1 Nuclear Accident on YouTube History Channel
  45. ^ Mahaffey, James (2010). Atomic Awakening. Pegasus Books. ISBN 978-1605982038.
  46. ^ Shulman, Review by Seth (19 April 2009). "Book Reviews: 'The Day We Lost the H-Bomb' - 'Atomic America'; by Barbara Moran - by Todd Tucker" – via

External links[edit]