Cruising roughly 300 miles northwest of Spain, the November-class nuclear submarine K-8 (see 13 Oct 1960) is taking part in a large Soviet naval exercise known as "Okean 70".

On 8 April, fires break out in compartments 8 & 9, asphyxiating 19 men. The sub surfaces so the crew can fight the fires. Automatic systems have shut down the reactor, so radiation is not a problem. However, the emergency generators cannot be started, so there is no electrical power to pump air into the ballast tanks. The stored compressed air that is keeping K-8 afloat runs out on 10 April. By this time, surface ships are standing by. U.S. Navy P3 antisubmarine patrol planes observe the K-8 lying dead in the water, with crew members on deck trying to rig a tow line to one of the ships. Late in the day, most of the crew transfers to these ships. But Moscow orders them back aboard to save K-8.

K-8 is taken under tow, but the tow cable snaps in heavy seas. At 0620 the next day, the stricken submarine sinks in 4,680 meters of water. Her captain and 22 crewmen go to the bottom with her. When the P3s arrive during full daylight, they find nothing but an oil slick at the location.

Soviet vessels stand sentinel duty over the K-8 for at least 6 months, and regular patrols visit the site until 1979. It is presumed this is to keep western military forces from attempting salvage.

At Cumbria in the north of England, the Windscale plant is used to recover plutonium from scrap. Windscale is thought to be a well-run operation, with effective safety features. One of these is a deep trap: a loop of hose that drops 25 feet from the transfer vessel, then comes back up to feed solution to the next stage of the process. Its purpose is to prevent backflow. The differential is achieved by having the transfer vessel 8 feet above the top of the trap.

When a brief, relatively small excursion is detected in a transfer vessel, it presents a mystery. It is known that the amount of plutonium in the batch is too small for critical mass, so the theory is that precipitate must have somehow collected in the vessel. Disturbing any such sediments might precipitate another and greater excursion.

A 6-inch hole is drilled in the ceiling and used as access to open the vacuum line going into the transfer vessel. Inspection is done with a fiber-optic probe developed especially for this problem. The vessel is found to contain just liquid, but both aqueous and organic solutions are present. The trap is found to hold 39 liters of organic solution that has been there for several months and perhaps as long as 2 years. Through that time it has been leaching plutonium from the aqueous solutions that passed through; it now contains a total of 2.15kg. Although the source does not provide every detail, it is clear that this is just a "red herring": the true answer to the puzzle is that when some aqueous solution has poured into the transfer vessel, it forms an emulsion with the organic — much like the oil & vinegar you shake for your salad dressing. When the pouring stops, the emulsion spreads out in a layer, making a favorable geometry that is just above critical. Then, over 10 seconds or so, the immiscible liquids separate and the reaction stops.

Two people are in the plant at the time of the excursion. They receive doses of 2 rads and less than 1 rad.

A fire breaks out on the stern of the submarine tender USS Canopus at Holy Loch, Scotland. This vessel carries several nuclear weapons and is moored alongside two nuclear submarines: the USS Francis Scott Key (SSBN-657) and the USS James K. Polk (SSBN-645). The Francis Scott Key casts off, but the Polk remains alongside.

U.S. naval authorities in Holy Loch and London reassure the public: "We have drills and precautions which rule out any danger whatsoever." It takes four hours to bring the fire under control; its cause is never determined. Three men are killed. U.S. Navy documents record extensive damage in the area of the fire.

A series of reflector-evaluation experiments at the Kurchatov Institute uses fuel rods enriched to about 20% U-235 (a value typical of icebreaker reactors.) Conducted at the critical experiment facility, SF-7, these experiments involve comparing iron and beryllium as reflectors. Fuel-rod configurations are variable, with a fixed lattice of boron carbide safety rods. Criticality is achieved by flooding the core with light water and regulating the neutron flux with the safety rods.

On the day of the accident, a new core configuration has been assembled and left overnight in the dry tank. The beryllium reflector is left in place because it was used in the previous phase of testing. However, calculations for this configuration have been done only for an iron reflector. Without waiting for the supervising physicist or the control console operator, the supervisor switches on the pump to begin filling the tank with water. Some instrumentation is powered on, but not all. In particular, the safety rods are not inserted.

As the tank fills, the supervisor begins talking with a scientist from Gorky who just arrived. Standing by the tank, they see a blue glow reflected from the ceiling and hear an audible neutron alarm. They think the trouble is elsewhere and race from the room. Other workers also leave. Later, a manager is informed and tries to enter the room. But radiation levels and steam from the tank make it impossible to approach the control console. After 5 to 7 minutes, power to the pump is shut off at a substation.

Later assessment shows about 50 pulses of neutrons have occurred, each with a yield of 5x10¹⁷ fissions. Total yield is estimated as 2x10¹⁹. The scientist and the supervisor each receive roughly 1,500 rem on their feet. Other personnel are behind heavy shielding and receive much smaller doses. Because the rate of criticality insertion is low, no fuel rods are damaged and the room is not contaminated.

In the SL-3 facility at Kurchatov Institute, a long series of experiments to measure the critical mass of a certain type of 90%-enriched fuel rod. The apparatus is broadly similar to that of the February incident: a large number of fuel rods of adjustable configuration, housed in a tank which is filled with water as a moderator. The water can be drained by a slow outlet in 15-20 minutes, or by a fast dump in 20 to 30 seconds. The crucial difference is that here the rods are supported by relatively fragile plexiglas plates.

After the final experiment concludes, the supervisor orders all control and safety rods inserted. The neutron source is removed from the core. Four staff members (including the supervisor) enter the room to examine the experiment. The supervisor then orders a fast dump of the water. The slow drain has been used in all previous experiments. Water pressure causes the plexiglas baseplate to deform; the fuel rods, unsupported, fall into a highly supercritical arrangement. The energy release, estimated as 5x10¹⁸ fissions, destroys some of the fuel rods and splashes water out of the tank.

Despite this, radioactive contamination of the room is minimal, and none escapes to the outer premises. However, all four personnel are severely exposed. A technician beside the tank receives 6,000 rem and dies in 5 days. The supervisor gets 2,000 rem and dies within 15 days. Doctors are able to save the other two, who receive doses of 700-800 rem, but both suffer long-term health impairment.

Karen Silkwood is a metallography laboratory technician at the Cimarron River plutonium plant of Kerr-McGee Nuclear Corporation in Crescent, OK. She joins the Oil, Chemical & Atomic Workers Union local at the plant and participates in a strike. That strike reduces the number of union members from 150 to 20, and Kerr-McGee schedules a decertification election. Silkwood is elected to the union bargaining committee with the assignment to investigate health and safety issues. In the summer of 1974, she testifies to the AEC that the company is careless with the plutonium it is building into reactor fuel rods. Safety standards allegedly have slipped because of a production speedup, which results in employees being given tasks for which they are poorly trained. Silkwood also alleges that Kerr-McGee welds the fuel rods improperly and falsifies inspection records.

In October or November of 1974, Silkwood is discovered to be significantly contaminated with plutonium. Strangely, though the gloves in the glovebox she uses to make the plutonium pellets are in good shape, there is contamination on their outside surfaces. Silkwood is decontaminated at the plant, but the next day is found to show alpha activity again, although she does not work at the glovebox that morning. Her locker and automobile show no activity. Her apartment, however, shows higher levels of alpha emission than the areas where she works. The worst is 400,000 dpm (disintegrations per minute) on a package of bologna and cheese in the refrigerator. Kerr-McGee arranges for her to be examined at Los Alamos, where a doctor assures her that the amount of plutonium in her body is not dangerous. The situation causes her some stress nevertheless. Under prescription, she begins using Methaqualone to calm her nerves.

Silkwood says she has documents to prove her allegations. Shortly after six o'clock on 13 November she leaves a union meeting at the Hub Cafe in Crescent. Another attendee of that meeting later testifies that she does have a binder and a packet of documents at the cafe. She gets into her car and heads alone for Oklahoma City, about 30 miles away, to meet with New York Times reporter David Burnham and Steve Wodka, an official of her union's national. She never gets to the meeting. Later that evening, her body is found in her car, which has run off the road and struck a culvert. The car contains no documents. There is no firm evidence of foul play, and the coroner finds 0.35 milligrams of methaqualone (Quaalude) per 100 milliliters of blood at the time of her death. That amount is almost twice the recommended dosage for inducing drowsiness.

Kerr-McGee closes its nuclear fuel plants in 1975. The grounds of the Cimarron plant are still being decontaminated 25 years later. In 1979, Silkwood's estate is awarded $10.5 million for personal injury and punitive damages, but this judgement is reversed on appeal. In 1986, the Supreme Court restores the original verdict. The suit is headed for retrial when Kerr-McGee settles out of court for $1.38 million, admitting no liability. Many aspects of the case are controversial. There are those who think Karen Silkwood was murdered on Kerr-McGee's orders. Others believe she stole some plutonium and deliberately contaminated herself, seeking to discredit the company. The BBC (which has the best summary) says "That controversy continues to this day. It seems likely that the facts will never be publicly known."

At the Radiation Technology, Inc. facility in Rockaway, New Jersey, Michael Pierson is exposed to 150-300 rems when a safety interlock system designed to protect workers from a cobalt-60 source fails.

In 1986, the NRC cites company executives for intentionally disabling the system. In 1988, after more than 30 NRC violations, company president Martin Welt and nuclear engineer William Jouris are charged in federal court with 11 counts of conspiracy to defraud the NRC, making false statements and violating the Atomic Energy Act. Welt, who threatened to fire workers who didn't lie to NRC investigators, is also charged with obstruction of justice. Both men are convicted. Joris is sentenced to probation; Welt is sentenced to two years in prison, placed on three years probation, and fined $50,000.

While workers are draining a piping system aboard the USS Puffer (SSN-562), radioactive water spills on the drydock surface at the Puget Sound Naval Shipyard in Bremerton, WA. A Navy spokesman says that "less than 5 gallons" of slightly radioactive water from the ship's secondary cooling system were lost, and that it spilled as the workers were draining the water into two 5-gallon containers, a routine procedure. According to the Navy, no workers were contaminated and the drydock drain was closed before any spillage escaped into the sea.

Shipyard employees reportedly dispute the Navy's account, saying that the spill was much bigger, about 100 gallons, that response to the spill was slow, and that several workers suffered skin contamination. A 15- by 20-foot section of drydock is jackhammered up, sealed in drums and shipped to a nuclear waste site in Hanford, WA.

A shielded facility at the Idaho Chemical Processing Plant recovers enriched uranium from solutions of spent fuel rods by interacting aqueous and organic solutions in columns. The operation is highly automated and rather complex, with various kinds of solutions interacting in several columns and connecting tubes in order to separate uranium from fission products while conserving resources and preventing high concentrations of radioactives to build up in any solution.

The accident has multiple causes. A leaky valve causes unnoticed dilution of a buffer solution, so that it takes up more uranium than expected. The tank density alarm that should indicate this is inoperative, and another tank has not had its density alarm installed. A chart recorder shows the departure from constant liquid level, but the leak is so slow that this would not be noticed unless specifically looked for. (In any case, the chart paper runs out weeks before the accident and is not replaced until afterward.) Finally, procedures requiring sample analysis are not being followed.

The effect of all these factors is to allow a gradual closed-loop concentration of uranium nitrate in two of the process columns. When radiation alarms begin sounding, the supervisor goes outside the plant and measures dose rates of 100 mRem/hour. He orders the building evacuated. This is accomplished in three minutes. No significant exposures result, and the estimated yield of 2.7x10¹⁸ fissions does not damage any equipment or contaminate the facility. The incident leads to an extensive review of procedures, better training and documentation, and installation of more instrumentation.

This facility at the Siberian Chemical Combine is where plutonium metal ingots are checked for purity and packaged for shipment onward. The administrative procedure calls for only two ingots per container. (The containers are lined with cadmium, so that placing multiple containers in close proximity is not a problem.)

On this occasion, production pressure apparently causes two workers to violate procedures by assisting each other. The end result is that one attempts to place a fourth ingot into a container already holding three. This fourth ingot is either violently ejected by the prompt heat of the criticality, or withdrawn by the operator when he sees the flash. In any case, the reaction is ended after triggering radiation alarms to cause evacuation of two buildings.

Later analysis of the ingots involved leads to a yield estimate of 3x10¹⁵ fissions. The equipment is not damaged and no contamination results. However, the operator who held the fourth ingot gets a whole-body dose of 250 rad and more than 2,000 rad on his hands. His arms are amputated up to the elbows, and he later has eye trouble as well. Seven others receive doses of 5 to 60 rad, mostly from fast neutrons.

Poor equipment design, component failures, and worker mistakes contribute to a partial meltdown of the core of reactor #2 at the Three Mile Island power plant near Middletown, Pennsylvania. Contaminated cooling water escapes into a nearby building, releasing radioactive gases that spread over the region. Over 200,000 people flee the immediate area, and what's known at "White-knuckle week at Three Mile Island" ensues as the event becomes a media circus, and the nuclear industry tries to regroup as protesters call for its abolition.

Metropolitan Edison, the operator of the plant, is indicted in 1983 for falsifying leak rate data at TMI-2 and for destroying documents before the accident. In 1984, it pleads guilty to one count and no contest to six counts of the 11-count indictment.

Investigation of the wreckage is done by remote control, and very cautiously. By 1985, robotic cameras show much greater damage to the core than expected. Temperatures reached 5,000 degrees; half the fuel elements melted and slumped down to the bottom of the containment, leaving a great void. The cleanup takes ten years and costs over a billion dollars. Based on this cost, and the amount of radiation released, TMI is the worst commercial nuclear reactor accident in the United States. It brings about wrenching changes in the way nuclear plants are built and operated, and it essentially stops any such plants being built in the U.S.

Still, health impacts of the radiation release remain controversial. The Rogovin Report is quoted as stating that "approximately 2.5 million curies of radioactive noble gases and 15 curies of radioiodines were released. These releases resulted in an average dose of 1.4 mrem to the approximately two million people in the site area." The majority of epidemiological studies performed since the accident show no correlation between this release and cancer or other adverse health impacts in the Susquehanna Valley. A study by Dr. Ernest J. Sternglass, professor of radiation physics at the University of Pittsburgh, claims that the accident leads to a minimum of 430 infant deaths. There is considerable justification for the industry's position that "no one died at Three Mile Island". In a lawsuit by 2,000 neighbors of the plant alleging radiation-induced injuries, a court finds no substantial proof of any damages. A 1994 survey by the Union of Concerned Scientists notes a paucity of investigation of this question, and predicts that solid answers may never be known.