"If the oceans were filled with liquid sodium, then some crazy scientist would want to build a water-cooled reactor.”

—Hyman Rickover, grousing about the sodium-cooled USS Seawolf

Admiral Hyman Rickover pushed his passion for a nuclear-powered submarine as hard as he could without being formally charged with criminal intent, and he was rewarded with one of the most successful projects in the history of engineering. His finished prototype submarine, the USS Nautilus, was all that he had hoped. First put to sea at 11:00 a. m. on January 17, 1955, she broke every existing record of submersible boat performance, made all anti-submarine tactics obsolete, and never endangered a crew member.

At the end of World War II the Navy, whether it knew it or not, was ready for the improved submarine power plant of a nuclear reactor. All the groundwork was in place after the Army’s startling success in the development of atomic bombs and the infrastructure for their manufacture. As an engineering exercise, the task of reducing the size of a power reactor from that of a two-story townhouse to something that would fit in a large sewer pipe was seen as possible, but there were a couple of serious considerations.

First, there was the nagging dread of a steam explosion. It was one thing to lose a boiler or two on a battleship, but in the tightly confined spaces of a submarine everyone would be killed and the boat lost if a steam line broke. If a reactor were to lapse into runaway mode, the water coolant could flash into high-pressure steam and tear the sub in half under water. Experience was in short supply and it was hard to convince mechanical engineers that this was a very unlikely occurrence in those early years of nuclear power. The use of water in the primary coolant loop was considered too dangerous to pursue.

Second, in the early 1950s there was concern over the availability of uranium as reactor fuel. To build exactly one uranium bomb in World War II the United States depended on all the uranium ore that could be mined in the Belgian Congo and Canada. There was no guarantee that enough uranium would ever be available to power a fleet of submarines, even from multiple foreign sources working at maximum efficiency. Plutonium, on the other hand, was a perfectly good power reactor fuel and we had plenty of it. It was manufactured at the Hanford Works in Washington State. It was therefore obvious that a nuclear submarine should be fueled with plutonium, and to fission plutonium fast neutrons were optimal. It was best to have no neutron moderation in a plutonium reactor, and this meant that there could be no water coolant in the primary loop. The coolant would have to be something heavy and liquid, such as melted metallic sodium. It would run hot and thermally efficient at atmospheric pressure, putting no expansive stress on pipes and associated hardware.

Rickover strongly disagreed with this “fast reactor” philosophy. Unlike many engineers assigned to the nuclear submarine project, Rickover had paid his dues riding around in the

ocean in submarines S-9 and S-48, which were feeble death-traps built in the 1920s.134 He knew from miserable experience and alert observation that there is no such thing in a submarine as a pipe, tank, flange, or valve that will never leak. In fact, with the combined stresses of being crushed, twisted, hammered, vibrated, abused, and built by the lowest bidder, a submarine’s bilge ditch would slosh with a sickening mixture of sea water, diesel fuel, sweat, lubricating oil, salad dressing, vomit, battery acid, head overflow, and coffee. Anything in the boat that conducted or contained compressed air or any fluid was capable of leaking, regardless of how well it was welded together or tightened with threaded fasteners. The captain in an S-boat had to wear a raincoat when operating the periscope, and dampness covered every surface inside the craft. Running in cold water meant that standing in the control room you could peer down the centerline and lose sight of the end of the compartment in the fog.

The concept of cooling an engine with liquid sodium thus seemed wrongheaded to this experienced submariner. The slightest sodium leakage would react with water or with water vapor in the air, burning vigorously and leaving a highly corrosive, flesh-eating sodium hydroxide ash. Sailors could perish just from breathing the hot vapor. Furthermore, Rickover had confidence that when there is an attractive price put on a mineral, such as uranium ore, people will dig up the earth to find it. In time he would be proven correct on both issues, the difficulties of using sodium and the abundance of uranium. With the safety of the crew being his primary design factor, his uranium-fueled pressurized-water reactor became the standard for submarine propulsion and for most of the civilian nuclear power industry. With a well-trained and disciplined reactor operations crew, there was no safer way to generate power in a confined space.

There were several experimental liquid-metal cooled and breeder reactors built in the United States, beginning in 1945 with Clementine, and unforeseen problems with these exotic designs

were experienced. Never was a serious sodium leakage encountered.135

The United States Atomic Energy Commission (AEC), starting work on January 1, 1947, had among its tasks the job of persuading private enterprise to build nuclear power plants. It was a noble goal, born of some very long-term projections. It was seen, even back in the late 1940s, that civilization would require more and more electrical power, and we could not generate it by burning coal forever. Coal and oil were seen as limited resources that were not being regenerated, and there were a finite number of rivers left to be dammed. Wind, used to make power since the days of the Roman Empire, was seen as too feeble and unreliable to meet the growing power demand, and solar power was limited to hot-water heaters in Florida. A plausible power supply for the future was nuclear fission, and it was not too early to begin an experimental phase of development.

A stellar committee met to come up with five nuclear reactor concepts to be prototyped and tested as candidates for the standard civilian power reactor. Uranium, the fuel of choice for all military reactors, was not seen as plentiful. Even if a hundred mines were dug, it was just another commodity that we would run out of eventually. Therefore, the ideal civilian nuclear power reactor would be a breeder, which paradoxically produces more fuel than it burns. Furthermore, the danger of a steam explosion should be minimized by vigorous application of the engineering art.

There are two types of breeder reactor, the fast breeder and the thermal breeder. The fast breeder runs on plutonium-239. The thermal breeder runs on uranium-233. The first of the five prototype reactor programs to be funded by the AEC in June 1954 was a sodium-cooled thermal breeder named the Sodium Reactor Experiment, or the SRE. Winner of the contract for the project was the North American Aviation’s Santa Susana Field Laboratory, located in a rocky wilderness about 35 miles northwest of Los Angeles, California, called Simi Hills.

The 2,850-acre site was divided into four sections. Three of the sections were devoted to testing high-performance rocket engines and explosives, and the fourth was populated with exotic nuclear-reactor experiments. The lab was thus blessed with all the fun stuff of the era except above-ground nuclear weapon and aircraft ejection-seat tests, and it should have been located in Idaho instead of in a place that would experience a population explosion in the next few decades. As the years passed, the valley below would jam tight with Californians, some of whom would find fault with the Santa Susana lab for no good reason other than what it was doing.

Admittedly, over the decades Santa Susana saw its share of accidents. The hot lab facility at the lab was the largest in the United States at the time. Workers could take apart highly radioactive reactor fuel assemblies using robotics behind windows three feet thick, offering complete radiation protection. Every now and then a fire would break out behind the hot-cell glass, causing massive internal contamination. The cores in four experimental reactors on site strayed outside the operating envelope and melted. Highly toxic waste disposal was handled by shooting at the barrels at a safe distance with a rifle until they exploded, sending the contents high into the air and wafting away into the valley below. In July 1994 three workers were trying to test some rocket fuel catalyst and it exploded unexpectedly, killing two, seriously injuring the

third, destroying a steel rocket fuel test stand, and setting a 15-acre brush fire.136 In 2005 a wildfire swept through Simi Hills and burned everything flammable in its path.

By April 25, 1957, the SRE was up and running and soon providing 6.5 megawatts of electricity to the Moorpark community, using a generator courtesy of Southern California Edison. It was the first civilian nuclear power consumed in the United States. In November, Edward R. Murrow featured the SRE power plant on his See It Now program on CBS television.

By current standards, the SRE was an odd power-reactor design. It was to run on fissions caused by neutrons slowed down to thermal speed, and yet the coolant was to operate at atmospheric pressure. The moderator material was graphite, a stable, solid-state material. Preventing any chance of a graphite fire, the moderator was formed into columns with a

hexagonal cross-section and covered with gas-tight, pure zirconium.137 The coolant was liquid sodium. Liquid sodium, being a dense, heat-conductive metal, is a very efficient coolant, and under foreseeable operating conditions, it never boils or produces dangerous gas pressure, as could water. It does, however, absorb an occasional neutron in a non-productive way, to much the same extent as ordinary water. Although the ultimate purpose of the SRE was to begin development of a civilian thermal breeder, its first fuel loading would be metallic uranium, slightly enriched to make up for the neutron losses in the coolant. As the basic configuration was proven by experiments, thorium-232 breeding material and uranium-233 fuel would be

introduced later.138

The Sodium Reactor Experiment ran with a graphite moderator at very high temperature, giving an efficient means to make high — pressure steam. The coolant was liquid sodium, which would never boil away, could operate at atmospheric pressure, and would never react chemically with the graphite. It was a sound concept, but the problems of dealing with a flowing liquid metal that reacts explosively with water have plagued reactors with sodium-based coolants.

There was another good reason to use sodium instead of water as the coolant. In water — cooled graphite reactors, such as the plutonium production reactors at Hanford, the unintended loss of coolant in the reactor always improves the neutron population, and the reactivity of the pile increases. The power starts going up without human intention. Graphite is a near-perfect moderator material, and any action that reduces the amount of non-graphite in a reactor, including the formation of steam bubbles, is fission-favorable. There was no such worry if sodium, with a boiling point of 1,621°F, were used instead of water. With an intended coolant outlet temperature of 650°F at full power, bubble formation in the reactor was hardly a concern, and the reactor could operate at an ideal temperature for external steam production while running under no pressure at all.

It was early in the history of power reactor development, and there were few successful plans to draw on, so there were novelties in the SRE embodiment. At least one aspect of the plan, the sodium pumps, seemed sub-optimal. The EBR-I sodium-cooled breeder reactor in Idaho had been built back in 1951 using exceedingly clever magnetic induction pumps for the coolant. A sodium pump in this system was simply a modified section of pipe, having a copper electrode on either side of the inner channel. A direct current was applied to the electrodes with a static magnetic field running vertically through the pipe, and the electrically conductive liquid metal was dragged along through the pipe by the same induced force that turns the starter motor on a car. The pump had no moving parts. EBR-I reported no problems using induction pumps, but the EBR-I was producing a scant 200 kilowatts of electricity.

For the SRE, hot oil pumps used in gasoline refineries were used to push the sodium.139 A large electric motor, capable of moving molten sodium at 1,480 gallons per minute up a vertical pipe 60 feet high, turned a long steel shaft, ending in a turbo impeller in a tightly sealed metal case. A single, liquid-cooled ball bearing supported the working end of the shaft. The problem of keeping liquid sodium from leaking past the impeller and into the bearing was solved by modifying the end of the pump. The shaft was sealed with a ring of sodium, frozen solid in place by a separate cooling system. The coolant to be pumped into the seal could not, of course, be water, which would react enthusiastically with the sodium. It had to be a liquid that had zero trouble being next to sodium. Tetralin was chosen.

1,2,3,4-tetradhydronapthalene, or “tetralin,” is a solvent, similar to paint thinner you would buy at the hardware store, first synthesized by Auguste George Darzens in 1926. Its molecule is ten carbon atoms and a dozen hydrogens, looking like two benzene rings stuck together. It has no particular problem with sodium, and it evaporates at about 403°F. The tetralin was circulated through the sodium seal, keeping it solidified, in a continuous loop using two parallel evaporation coolers to shed the heat from the seal. Electrically driven pumps kept it moving, with a gasoline engine for a backup in case of an electrical failure. It was a complicated sub­system in a complicated power plant, requiring pipes, valves, pumps, wiring, instrumentation, tanks, and coolers, just to keep the sodium off the pump bearings. The fact that it had to be backed up was ominous. It is always a better system when if everything fails, the wreckage reduces to an inert, safe condition.

The cooling system used three closed loops. The primary loop was liquid sodium running through the reactor. The natural sodium was constantly being activated into radioactive sodium — 24 by contact with the neutrons in the reactor. To eliminate the potential accident of radioactive sodium leaking into the steam system, a second sodium loop took the heat from the first loop and took it outside the reactor building, where it was used to generate steam for the turbo­generator in a water loop. For low-power experiments in which electrical power was not generated, the water loop was diverted into an air-blast heat exchanger, dumping all the power into the atmosphere. There was no danger of broken or melted fuel leaking radioactive fission products into the surroundings, as the second sodium loop was a well-designed buffer. No expensive containment structure was needed for the reactor, because there was no chance of a radiation-scattering steam explosion in the building. The steam was generated out in the yard, and it was not connected directly to the reactor.

Radioactive gases produced by fission, such as vaporized iodine-131 and xenon-135, were controlled in the stainless steel reactor tank by a bellows structure at the top, giving the tightly sealed system room to expand. Helium was kept over the reactor core to prevent air from leaking in and reacting with the sodium. The fission gases were piped off and compressed into holding tanks for controlled release into the environment after having been held long enough for the radioactivity to have decayed away. There were outer space reactors, rocket engines, and military systems being developed at Santa Susana, and all had their considerations of performance, weight, and reliability, but this SRE was to be a prototype civilian power plant. As such, the prevention of harm to the public was a primary and noble design consideration.

All newly designed sub-systems were actively tested at Santa Susana before they were integrated into this new type of power reactor. It was an experimental setup stacked with many unknowns, and there was a lot to learn about graphite-moderated, sodium-cooled reactors. The operation log shows that almost immediately there was trouble with the sodium pumps. Hours after the first power startup on April 25, 1957, the shaft seal on the main primary pump failed. A few weeks later, tetralin was leaking from an auxiliary secondary pump. A week later, the main secondary pump was replaced. In August, the sodium was found to be contaminated with tetralin. In November, the cold trap was clogged with sodium oxide. Air was getting in somewhere. In January 1958, sodium smoke filled the high bay, and men had to go in with oxygen masks to find the leaking bellows valves. In May the shaft seal on an auxiliary primary pump failed, and out of a main primary pump they could smell the strong odor of tetralin. Two months later all the pumps were taken apart and the ball bearings were replaced. The entire main primary pump was replaced with a spare. A month later, the electromagnetic pump clogged with sodium oxide again, and in September the main secondary pump was leaking tetralin. By April 1959 the main primary pump was leaking tetralin from the sodium seal, and the entire unit had to be replaced. A month later, a tetralin fire was extinguished without causing

reportable damage.140

So, pumping sodium around in a nuclear reactor was not as easy as it seemed on paper, and the SRE was being taken apart and worked on all the time. That is the life of an experimental reactor, and Rickover’s insistence on no sodium in his precious Nautilus seemed to make sense if you were outside looking in. Every time a component from the reactor tank or the primary coolant loop was removed for repair or examination, the sodium frozen to it had to be cleaned off. For this purpose, the pump or the fuel assembly was moved to the wash cell, a special setup behind an atmospherically sealed hot-cell window. Using remote arms, a technician would hose down the sodium coating with warm water. It would instantly turn into hot sodium hydroxide and wash down the drain at the bottom of the stainless steel hot-cell chamber, leaving the once-contaminated piece bright and sparkling clean.

Real trouble did not start until RUN 13, from May 27, 1959 to June 3, 1959. The crew was supposed to run the coolant outlet temperature up to 1,000°F to see if the system could stand to work at higher power. They hoped to log 150 megawatt days. All was well until two days after startup at 11:24 a. m., when the reactor scrammed due to an abnormal sodium flow rate. Not hesitating to contemplate why the flow rate was wrong, the operating crew restarted the reactor immediately and ran it back up to power. There it stayed until 9:00 the next morning, May 30. At that point, the reactor system went squirrelly.

First, the reactor inlet temperature began to rise slowly over three days. On June 1, the temperature difference across the heat exchanger rose sharply, indicating that something wasn’t working. The thermocouple in one fuel assembly, number 67, showed a temperature increase from a normal 860°F to 945°F. The temperature in the graphite abruptly jumped by 30°F, also on May 30, and the thermocouple in fuel assembly number 16 showed a similar increase in temperature. They did not notice it at the time, but the automatic control-rod positioner was compensating for a slow increase in reactivity in the reactor. Obviously, something had occurred that was impairing the coolant flow, and by June 2 the main primary pump casing was reeking of leaking tetralin. The reactor was shut down on June 3 to examine the fuel and repair the coolant pump.

As the pump was being torn down, fuel assembly number 56 was removed using the impressive automatic fuel-removal machine and transferred to the wash cell for examination. To quote the accident report exactly, “During the washing operation a pressure excursion occurred of sufficient magnitude to sever the fuel hanger rod and lift the shield plug out of the wash cell.” Translation: The damned thing exploded and put the wash cell out of commission for a year.

Nobody was killed, thanks to the three-foot-thick window and aggressive ventilation.141

Retrospective analysis would find that the vent at the bottom of the fuel assembly, where coolant was supposed to flow in and past the hot fuel rods, had been blocked by a substance technically referred to as “black stuff,” and this left a large remainder of sodium in the bottom of the assembly. When the technician aimed the hose into the top of the assembly, the big sodium wad went off like a hand grenade. This was not noted at the time, and number 56 was put back in the reactor. Damage to the wash cell diverted attention from further identification of the black stuff, but a working explanation was that it was residue from tetralin decomposed in the hot coolant. There was not supposed to be any tetralin in the coolant, but three pints of it were found in the cold trap plus a couple of quarts of naphthalene crystals, or tetralin with the extra hydrogens stripped off. No connection in particular was seen between these contaminations and the strange behavior at the end of RUN 13. The troublesome tetralin-cooled seal on one pump, the main primary, was replaced with a nack-cooled sodium seal, and the reactor was

ready for RUN 14.142

RUN 14 was started on July 12, 1959. The experimenters expected some trouble with the fuel-channel outlet temperature, but they were not sure why. Perhaps if they could intensify the effect, the cause would snap into focus. The reactor was brought smoothly to criticality at 6:50 am. At 8:35 they increased the power level to a modest 500 kilowatts, and the graphite temperature started to flop around wildly, running up and down by about 10°F, and various fuel — channel temperatures started to diverge by 200°F. Not seeing this as a problem, they kept going until 11:42 when the reactor, hinting that something was amiss, scrammed due to a loss of sodium flow in the primary loop.

At this point I must find fault with the way they were operating the SRE. Something about the reactor was not working, and yet they kept restarting it without knowing why. Today, this disregard of trouble signs would be unheard of, I hope. You would never restart a reactor without knowing exactly why it scrammed. There would be inquiries, hearings, lost licenses, and firings, but apparently not in 1959, when the screws of bureaucracy weren’t tightened as they seem now. Start her up again and let’s see if that was just a fluke. By 12:15 they had SRE back up to power and were increasing the power and the outlet temperature. At 1.5 megawatts the temperature was fluctuating inexplicably by 30°F.

At 3:30 P. M., both air-radiation monitors in the reactor building indicated a sharp rise in activity. By 5:00 P. M., radioactive air was going up the exhaust stack and into the atmosphere, and the radiation level over coolant channel seven was extremely high, at 25 roentgens per hour. Something had obviously broken, but for some reason it did not occur to the experimenters that radioactive products that would cause such an indication are produced in the fuel, and the fuel is welded up tight in stainless steel tubes. If all the fuel is intact, then nothing radioactive should be leaking into a coolant channel, and certainly not through the gas-tight reactor vessel and into the air in the room. By 8:57 P. M. they had shut down the reactor. They fixed the problem of leaking radioisotopes by replacing the sodium-level indicator over channel seven with a shield plug, and they restarted the reactor at 4:40 am. the next day, July 13.

By 1:30 that afternoon, they noticed that the graphite temperature was not going down when they increased the coolant flow. They knew not why. At 5:28 P. M., they were running at 1.6 megawatts and commenced a controlled power increase. The power seemed to increase faster than one would expect up to 4.2 megawatts, but then the reactor went suddenly subcritical and the power was dropping away. They started pulling controls to bring it back, and by 6:21 P. M. they had managed to coax it to run at 3.0 megawatts.

Up to this point, the reactor had been recalcitrant and unusual, but now it went rogue. Power started to run away. Control-rod motion was quickly turned around, sending them back into the reactor to soak up neutrons and stop the power increase. Instead, the power rise speeded up. By 6:25 P. M., the power was on a 7.5-second period, or increasing by a factor of 2.7 every 7.5 seconds. The fission process was out of control, and a glance at the power meter showed 24 megawatts and climbing.

Deducing that if this trend continued, then in a few minutes the reactor would be a gurgling puddle in the floor, the operator palmed the scram button, throwing in all the controls in the reactor at once and bringing the errant fissions to a stop.

As the reactor cooled down, only one question came up: Why was there not an automatic scram when the reactor period, spiraling down out of control, passed 10 seconds? The period recorder, which leaves a blue-line-on-paper graph of period versus time for posterity, had a switch that was supposed to be tripped when the pen hit 10 seconds on the horizontal recorder

scale. This switch was supposed to trigger an automatic scram.143 Testing found that the switch would have worked, but only if the period had been falling slowly. The trip-cam was modified so that the switch would operate even when a scram was most needed, with the period falling rapidly.

Seeing nothing else to fix, the operating staff brought the reactor back to criticality at 7:55 P. M. and proceeded to increase the power. By 7:00 am. the next morning, July 14, they were running hot, straight, and normal at 4.0 megawatts. Two hours later, the radioactivity in the reactor building was reading at 14,000 counts per minute on the air monitors. Technicians put duct tape over places where fission products were found escaping. There was only one other automatic scram for the whole rest of the day, when workers setting up a test of the main primary sodium pump accidentally short-circuited something.

On July 15, it was seen as pointless to be trying to run the electrical generator while trying to test at high temperature, so the staff drained out the secondary coolant loop and switched to the air-blast heat exchanger. The next day, at 7:04 am., the SRE was made critical once more. On July 18, the motor-generator set, which was supposed to prevent power surges into the control room instruments, failed. The operators switched power to unstabilized house current and continued operation. At 2:10 a. m. on July 21, the reactor scrammed suddenly, having picked up another fast power rise. The scram was attributed to the unstabilized power, and the reactor was restarted 15 minutes later. One more scram at 9:45 am.

The next day, July 22, channel 55 was giving trouble. This assembly contained various experimental fuels, and the temperature was fluctuating in the 1,100 to 1,200°F range. There was only one automatic scram on July 23, probably just a fluke, but by 1:00 PM. the temperature in channel 55 was up to an eyebrow-raising 1,465°F. Operation continued. In the early morning on July 24, eight hours were spent trying to dislodge some apparent debris stuck in the fuel channels by jiggling the assemblies. It was noted in passing that four of the fuel modules seemed jammed and stuck firmly in place. There were two annoying automatic scrams later that day.

At 11:20 am., the Sodium Reactor Experiment RUN 14 was terminated, and the long-suffering machinery was allowed to rest quietly while the experimenters poked around in the core with a television camera. To their surprise, they found that the core of a nuclear reactor that had been acting oddly for six weeks, subjected to overheating and temperature fluctuations, many automatic scrams, pump seal coolant failures, oxidizing sodium, radiation leakage, and a power runaway, was wrecked. Of the 43 sealed, stainless-steel fuel rods in the core, 13 had fallen apart, scattering loose fuel into the bottom of the reactor vessel. How the thing had managed to run at all under this condition was an amazement in itself. Attempts to remove the fuel rods came to an end when the contents of channel 12 became firmly jammed in the fuel handling cask. An investigation of the damage and its cause was started immediately.

The interim report, “SRE Fuel Element Damage,” was issued on November 15, 1959. It was found that tetralin had been leaking into the sodium coolant through the frozen sodium seals in the pumps. In the high-temperature environment of the active reactor core, the solvent had decomposed into a hard, black substance, which would tend to stick in the lower inlet nozzles of the graphite/fuel modules and prevent coolant from flowing. In the blocked coolant channels, the sodium vaporized, which had been thought unlikely, and denied coolant to the fuel. The stainless steel covering the cylindrical fuel slugs melted, and structural integrity of the fuel assemblies was lost. Naked uranium fuel, having fissioned for hundreds of megawatt-hours, was able to mix with the coolant. Gaseous fission products presumably escaped the sealed reactor vessel, probably through the same leakpoint that allowed the sodium to oxidize, and other fission waste dissolved in the coolant, making the primary loop radioactive.

The most puzzling part of these findings was: Why did the stainless steel melt instead of the metallic uranium? The 304 stainless used for fuel cladding melts at 2,642°F, and the temperature in the reactor was nowhere close to that. Uranium melts at 2,060°F, and yet the stainless steel melted away, leaving the uranium unclad and unsupported. The fuel assemblies were, however, operating at temperatures for which the reactor was not designed. Where fuel slugs within the tubes were leaning against the inner walls, uranium diffused into the stainless steel, making a new alloy, a stainless steel/uranium eutectic. This uranium/steel mixture melts at 1,340°F, so with the reactor core overheating, the fuel assemblies fell apart. This problem would have been hard to foresee.

Another puzzle was harder to figure out. Why did the reactor run away, increasing power on a short period? The power excursion was simulated using the AIREK generalized reactor-kinetics code running on an IBM 704 mainframe. If the parameters were tweaked hard enough, the simulation could even be forced to agree with the recorded data, but even a well-tuned digital simulation could not indicate why the transient had occurred. It was easier to explain the subcritical plunge than the short-period lift-off. Further calculations and physical experiments proved that it was the sodium void in the blocked coolant channels that caused the fission process to run wild. It was the same phenomenon that caused the fuel assemblies to melt. Graphite is a better neutron moderator than just about anything, and the fission process improves when there is no other substance, such as coolant, in the way.

Those inert, radioactive gases produced in the accident that had not leaked out and made it up the exhaust stack, xenon-135 and krypton-85, were kept in holding tanks for a few weeks for the radiation to decay away and then slowly released up the stack and into the environment. The iodine isotopes had apparently reacted with something in the building and were not found in the released gas. On August 29, 1959, a news release was issued to the Associated Press, United Press International, The Wall Street Journal, and seven local newspapers, informing the public of the incident. The wording tended to downplay it to the point that one would wonder what made it news, beginning with “During inspection of the fuel elements on July 26 … a parted fuel element was observed.” So? What’s a “parted fuel element”? No big deal was made.

The SRE fuel melt was a unique accident, and yet it was typical. It was unique in that so many obvious trouble clues were ignored for so long. I am not aware of another incident quite like this. It was typical of reactor accidents in that nobody was hurt, and the only way you could tell from looking at it that something had happened was to take it apart. As in many cases of reactor accidents, the fuel was damaged by a lack of coolant. There was no steam explosion. Lessons were learned, extensive modifications were made to the system to improve its reliability. For the newly funded SRE Power Expansion Program, the hot-oil pumps were junked and Hallum-type primary and secondary sodium pumps were installed, and SRE was restarted

on September 7, I960.144 It ran well, generating 37 gigawatt-hours of electricity for the Moore Park community. It last ran on February 15, 1964, and decommissioning of the nuclear components was started in 1976. In 1999, the last remainder of the Sodium Reactor Experiment was cleaned off Simi Hills. The idea of a graphite-sodium reactor died.

Real trouble did not begin until February 2004, when locals filed a class action lawsuit against the Santa Susana Field Laboratory’s current owner, The Boeing Company, for causing harm to people with the Sodium Reactor Experiment. They had been stirred to action by a new analysis of the 1959 incident by Dr. Arjun Makhijani, an electrical engineer. In a way, the suit was like building a house in the glide path of an airport and then suing because airplanes were found to be flying low overhead. When Santa Susana was built, Simi Valley was a dry, desert-like landscape. Makhijani estimated that the accident released 260 times more iodine-131 than the Three Mile Island core melt in Pennsylvania in 1979, which speaks well of Three Mile Island. His estimate was speculative, because no iodine-131 contamination could be detected at the time. Over 99% of the volatile isotope was captured as it bubbled up out of the naked fuel into the coolant, becoming solid sodium iodide, which collected in the cold trap in the primary cooling loop. Any pollutant that might have made its way up the stack was diluted in the air, and 80 days later there could be no detectable trace, as if there were any to begin with. There were no milk cows living in Simi Valley and no edible grass to contaminate. There was no detectable thyroid cancer epidemic. Nobody was hurt. Boeing settled with a large payout to nearby residents.

In those early decades of nuclear power, it was an unwritten rule in the AEC that the public was not to be burdened with radiation release figures or the mention of minor contamination. It was true that the general population had no training in nuclear physics and radiation effects, and if given numbers with error bars and a map of an airborne radiation plume, imaginations could take control in nonproductive ways. Nobody wanted to cause a panic or unwarranted anguish or to undermine the public’s fragile confidence in government-sponsored research. The results of such a policy are worse than what it is trying to forestall, as the government is commonly accused of purposefully withholding information, and misinformation rushes in to fill the vacuum. Conspiracy theories thrive. This fundamental problem of nuclear work has yet to be turned around.

Our next adventure in sodium takes us to Lagoona Beach, which sounds like a secluded spot somewhere in Hawaii, but it’s not. It is in Frenchtown Charter Township, Michigan, 27.8 miles from downtown Detroit, looking out onto Lake Erie.

Walker Lee Cisler was born on October 8, 1897, in Marietta, Ohio. An exceptionally bright student, Cisler sealed his fate by receiving an engineering degree at Cornell University in 1922. With that credential, he was hired at the Public Service Electric and Gas Company in New Jersey, was named chief of the Equipment Production branch of the U. S. War Production Board in 1941, and in 1943 was tapped as the chief engineer for Detroit Edison.

In 1944 he joined the Supreme Headquarters, Allied Expeditionary Force (SHAEF) in Europe, assigned the task of rebuilding the electrical power systems on the continent as the German army retreated. By 1945, he had the power system in France generating more electricity than it had before the war. Impressed, the early embodiment of the AEC named him secretary of the AEC Industrial Advisory group in 1947. Resuming work at Detroit Edison, he became president in 1951 and CEO in 1954.

As a visionary, pushing for a greater and everlasting energy supply in his native land, Cisler became an early advocate of the breeder reactor concept, and by October 1952 he established the Nuclear Power Development Department at Detroit Edison. His dream of a civilian-owned, commercial breeder played right into the AEC plans, and it was the second breeder concept in their set of demonstration power reactors to be built. The first, the thermal U-233 breeder, would be built at Santa Susanna. The second, the fast plutonium breeder, would be Cisler’s baby. A kick-off meeting was held with the AEC at Detroit Edison on November 10, 1954. Present at the meeting was Walter Zinn, the scientist who headed the slightly melted EBR-I project at National Reactor Testing Station in Idaho. The plant would be named Fermi 1, in honor of the man whose name, along with Leo Szilard’s, was on the patent for the original nuclear reactor. Ground was broken at Lagoona Beach in 1956, after Cisler had secured $5 million in equipment and design work from the AEC and a $50 million commitment from Detroit Edison.

It would be a long crawl to implementation of Cisler’s plan, fraught with ballooning costs, many engineering novelties, and strong opposition to the project by Walter Reuther. Reuther was an interesting fellow. A card-carrying Socialist Party member, anti-Stalinist, and a fine tool & die machinist, he became a United Auto Workers organizer/hell-raiser and was attacked by a phalanx of Ford Motor Company security personnel in the “Battle of the Overpass” in 1937.

This, at the very least, made him a well-known figure in Detroit.145 Reuther, the UAW, and eventually the AFL-CIO filed suit after suit opposed to the building permit for the plant and later the operating license, based on multiple safety concerns and the fact that it was not an automobile. The suits ate up a vast amount of time and money, and court decisions finding against Fermi 1 were taken all the way to the Supreme Court. In the summer of 1961, the court decided seven to two in favor of Cisler, the AEC, and Detroit Edison. Construction could proceed, and the projected cost had risen to $70 million.

Because Fermi 1 was designed as a plutonium breeder, liquid sodium was chosen as the coolant. A fast breeder core was much smaller than a thermal reactor of similar power, so the coolant had to be more efficient than water, and to hit the activation cross section resonance in uranium-238, the breeding material, the fast neutrons from fission had to endure a minimum loss of speed. Sodium was the logical choice, and two sodium loops in series were used, similar to the SRE design. It doubled the complexity of the cooling system, but it ensured that no radioactive coolant could contaminate the turbo-generator.

The use of sodium made things difficult and complicated. Refueling, for example, was comparatively simple in a water-cooled reactor. To refuel a water-cooled reactor you shut it down, and after it cools a while, unbolt the vessel head, crane it off, and lay it aside. Flood the floor with water, which acts as a coolant and a radiation shield. You pick up the worn-out uranium in the core, one bundle at a time using the overhead crane, and trolley it over to the spent-fuel pool. Carefully lower new fuel bundles into the core, drain the water off the floor, and replace the steel dome atop the reactor vessel. You’re done.

There is no way to do this with a sodium-cooled system. For one thing, the sodium gets radioactive running in the neutron-rich environment of fission, and it would take a week of down­time just to let it decay to a level of marginal safety. You cannot let it cool down, because in this state the sodium is solid metal, and you could not pull the fuel out of it. It cannot be exposed to air, and everything done must be in an inert-gas atmosphere. The entire building would have to be pumped down and back-filled with argon. Also, sodium, liquid or solid, is absolutely opaque. You cannot see through it, so, unlike looking down from the refueling crane through the water to see what you are doing, you would have to be able to refuel the reactor blindfolded. If anything weird happened in a water-cooled reactor core, you could conceivably stick a periscope down in it and have a look. Not so with sodium coolant. Fermi 1 was to be an electrical power-production reactor, running at 200 megawatts. Although in later terms this would be a very small reactor, in 1961 it was a clear challenge to design a machine of this size running in sodium.

The problems of dealing with refueling a sodium-cooled fast reactor were solved by enclosing the works in a big, gas-tight stainless steel cylinder, about two stories tall. In the bottom of the cylinder, off to the side, was an open-topped stainless tank containing the U-235 reactor core,

surrounded by the U-238 breeding blanket.146 Liquid sodium in the primary cooling loop was pumped into the bottom of the reactor tank through a 14-inch pipe. The sodium flowed up, through small passages between fuel-rods, taking the heat away from the fissioning fuel. The thick, hot fluid was removed by a 30-inch pipe at the top and sent to a steam generator for running the electrical turbo-generator. A “hold-down plate” pressed down on the top of the fuel to keep it from being blown out of the can by the blast of liquid sodium coming up from the bottom. Under normal operation, the liquid sodium in the structure was 34 feet deep. Any open space left was filled with inert argon gas.

An ingenious, complex mechanism was used to refuel the reactor and extract and replenish the irradiated U-238 rods in the breeding blanket. To one side of the reactor tank and inside the stainless vessel was another round can, this one containing a revolving turntable loaded with fuel and blanket rod-assemblies. It worked like the cylinder in a revolving pistol or the carousel in a CD changer, indexing around and stopping automatically with a rod assembly in position to be extracted or inserted. An exit tube led vertically to an airlock on the top floor of the reactor building, where a refueling car ran on rails and was able to reach down into the reactor vessel, through the air lock, and remove or replace rod assemblies in the rotor. The car, gamma-ray shielded by 17,500 pounds of depleted uranium, was driven back and forth on the rails by a human operator, under absolutely no danger from radiation in the reactor, the coolant, or in fuel being inserted or removed.

To refuel the core, the hold-down device was first lifted off the top of the core. The fuel — transfer rotor was turned by a motor until it clicked into place with an empty slot under the handling mechanism, which was a long, motor-driven arm hanging down next to the reactor. The arm then turned to the core, gently snatching a designated rod assembly, picking it up and out of the reactor, swinging over to the transfer rotor, and lowering it down into the open slot. The rotor then rotated clockwise and clicked to the next open slot. After the rotor became completely full, the refueling car would bring up the rod assemblies one at a time and put them into shielded storage. To put new fuel or breeding material in the reactor, the process was reversed, with the empty rotor first loaded with fresh fuel from the refueling car. The rotor would index around, presenting the new rod assemblies one at a time to the arm, which would pick them up, swing them through the core, and insert them. All this happened under liquid sodium. It was a totally blind operation, depending only on mechanical precision to find rod assemblies where they were supposed to be and transporting them with great accuracy. When a refueling operation was completed, the hold-down device would descend and cover the top of the core. To be able to automatically refuel this reactor from the comfort of a swivel chair by

pushing a few buttons was a marvel to behold.147

Given the recent tribulations at the SRE, great attention was given to the design of the coolant pumps. Made of pure 304 stainless steel, one primary pump was the size of a Buick. It was a simple centrifugal impeller, driven by a 1,000-horsepower wound-rotor electric motor, with a vertically mounted drive shaft 18 feet long. It turned at 900 RPM to move 11,800 gallons of thick, heavy, 1,000°F liquid sodium per minute and lift it 360 feet. The motor speed was controlled using a simple 19th century invention, a liquid rheostat, as once used to dim the lights in theaters.

The Fermi-1 fast breeder reactor was a bold move in the development of commercial power reactors, ensuring a future with no problems of obtaining fuel to generate power in a growing economy. Unfortunately, the fuel-supply problem failed to materialize, and the complexity of its sodium-based cooling system proved to be its downfall.

There were no liquid-cooled ball bearings to worry about, and no seal to keep sodium out of the pump motor. There was no tetralin seal coolant, of course. The main bearing at the impeller end was a simple metal sleeve, with the stainless shaft running loose inside it. It was a “hydrostatic bearing,” like the bearing on a grocery-cart wheel, but instead of having oily grease to keep it from squeaking, the lubricant was liquid sodium, supplied out of the fluid that the impeller was supposed to pump. The sodium was allowed to puddle up in the long shaft gallery, rising several feet above the impeller housing. Its depth was controlled by compressed argon gas introduced into the top of the pump, and this kept the sodium off the motor and out of contact with air.

As one last step before the reactor tank was assembled, the engineers decided to enhance a safety feature. A problem with fast reactors was that in the unlikely event of a core meltdown, the destroyed fuel matrix, becoming a blob of melted uranium in the bottom of the tank, could become supercritical. In this condition, it would continue making heat at an increasing rate and exacerbating a terrible situation. An ordinary thermal, water-cooled reactor, losing its symmetry and carefully planned geometric shape in a meltdown, would definitely go subcritical, shutting completely down. The thermal reactor depends greatly on its moderator, the water running

between fuel rods, to make fission possible. The fast reactor does not.148 To ensure a subcritical melt in a fast reactor, at the bottom of the tank is a cone-shaped “core spreader,” intended to make the liquefied fuel flow out into a shape that does not encourage fission, flat and neutron-leaky, like pancake dough in a skillet. Before the core tank was finished, the builders installed triangular sheets of zirconium on the stainless steel core spreader, just to make it more high-temperature resistant. They beat it into shape with hammers, making the sheet metal conform to the cone. This last-minute improvement was not noted on the prints approved of by the AEC.

During operational tests, the check valves on the pumps, meant to prevent backflow if something stopped working, would slam shut instead of closing gently but firmly, as they were supposed to. This flaw was re-engineered, and further extensive testing would ensure no problems from the pumps.

Still, working with a large volume of pure sodium would prove challenging. Tests were being performed in an abandoned gravel pit about 20 miles north of the building site on August 24, 1959, when a load of sodium exploded in air. Houses in the nearest neighborhoods, Trenton and Riverview, were damaged by the blast and six people were hospitalized. By December 12, 1962, there was enough of the reactor assembled to test the main cooling loop. An operator was at the control console watching the instruments as the sodium circulated at full speed through the system. The temperature started reading high on a thermocouple gauge. There was no fuel in the reactor, so where was the heat coming from all of a sudden? Were sodium and water mixing due to a flaw in the steam generator? Thinking this through, the operator reached over and hit the red water-dump button. Water gushed out of the secondary cooling loop into a holding tank, taking it quickly out of the steam generator and away from possible contact with the sodium loop. Unfortunately, this caused a sudden vacuum in the system, which was designed to hold high-pressure steam and not an airless void. A safety disk blew open, and sodium started oozing out a relief vent, hit the air in the reactor building, and made a ghastly mess.

No one was hurt. When such a complicated system is built using so many new ideas and mechanisms, there will be unexpected turns, and this was one of them. The reactor was in a double-hulled stainless steel container, and it and the entire sodium loop were encased in a domed metal building, designed to remain sealed if a 500-pound box of TNT were exploded on the main floor. It was honestly felt that Detroit was not in danger, no matter what happened.

Cost of the Fermi 1 project reached $100 million, and it was too far along to turn back. The

fuel was loaded on July 13, 1963, and the fuel car was not acting well.149 The first startup was a few weeks later, on August 23 at 12:35 P. M. A system shakedown at low power would continue until June of 1964. A few problems surfaced. The number 4 control rod delatched from the drive mechanism, leaving it stuck in the core. The large, rotating plug in the top of the reactor vessel, used to move the refueling arm between the fuel rotor and the core, jammed and wouldn’t move. A sodium pump had to be repaired, and the cap on the reactor vessel had to be rebuilt so that it would fit correctly. Some electrical connections and cable runs were defective, causing instrument problems. These glitches were all knocked down.

By January 1966, the Fermi 1 plant was wrung out and ready to go to full 200 megawatts of heat in several cautious steps. August 6, 1966, was a day of triumph. The thermal power was brought up to 100 megawatts, enough to make 33 megawatts of electricity, or about half what the backup diesel generators could produce. The project cost had also reached a high point, at a cool $120 million, and critics pointed out that the reactor had so far been able to generate measurable electricity for a total of only 52 hours.

At this half-power level, unusually high temperatures were indicated in fuel assemblies M-091 and M-140, the steam generator started leaking steam, and control rod no. 3 seemed to stick in the guides. The next day, August 7, the positions of the hot fuel assemblies were swapped with trouble-free fuel assemblies, to see if the problem moved. There seemed no correlation between the specific fuel assemblies and overheating. The problem seemed to be the position

in the core, and not the fuel.150 Could it be that the thermocouples in those locations were just reading wrong?

The operating crew was ready to try another cautious power-up on October 4, 1966. The Fermi 1 reached criticality at 11:08 P. M., and it idled at low power while every little thing was checked. At 8:00 am. the next day, the operators were ready to bring the reactor to half-power, but a steam-generator valve seemed stuck. It took until 2:00 P. M. to resolve that problem, and then the feed-water pump in the secondary loop was not working. They powered down and worked on it. By 3:05 P. M. they had resolved the problem and the power was increased to 34 megawatts and rising.

Something was not right. The neutron activity in the reactor core was erratic and bouncy. There was no reason for the neutron level to be anything but smooth and steady. The power ascension was halted while Mike Wilbur, the assistant nuclear engineer in the control room, contemplated the meaning of these instrument readings. Based on previous problems, Wilbur had a hunch. He stepped behind the main control panel to take a look at the thermocouple readouts on the fuel assembly outlet nozzles. These instruments were not considered essential for running the power plant, so they were not mounted on the main panels. They were included in the hundreds of instrument readouts, lights, and switches in the control room for diagnostics, and here was a problem that required a deeper look.

Fuel assemblies M-140 and M-098 were both running hot. At this power level, the temperature of coolant flowing out the top of fuel assembly M-140, which had given trouble in the past, should have been 580°F. It was reading 700°F. As Wilbur was taking this in, at 3:08 P. M. the building radiation alarms started sounding. It was a rude, air-horn sound: two mind — numbing blasts every three seconds. There were several possible explanations for the radiation alarms, but the assistant nuclear engineer knew deep in his heart that one was likely. Fuel had melted, spreading fission products into the coolant. The only thing that was not clear at all was: Why?

The crew executed emergency procedures as specified in the operations manual. All doors were closed, and all fresh-air intakes were closed in the building. Detecting radiation in the building was an emergency condition. It was supremely important to not let it leak out into the world outside Lagoona Beach, which would make it a big, public emergency. They executed a manual scram at 3:20 P. M., shutting Fermi 1 down with the floor-trembling shudder of all controls dropped in at once. One rod would not go in all the way. Not good. Was a fuel assembly warped? They tried another scram. This time, it went in. With the neutron-poisoning control rods all in, there was no fear of the core being jostled or melted into a critical condition. There could be no supercritical runaway accident, and if the core were completely collapsed and flowing onto the spreader cone, the uranium would be mixed with melted control material, which would definitely discourage fission. News of the accident, specifying that the engineers did not know what had happened, spread across the land.

The reactor had never run at full power, and only for a short time trying to get to half power, so there was no worry that it could melt down any further in the shutdown condition. There were not enough delayed fissions and fission product decays to cause havoc with high temperatures in the fuel. Over the next few weeks, the operators and engineers tried to find the extent of the damage without being able to see inside the reactor core. One at a time, they pulled control rods and noted the increase in neutron activity in the core. A few control rod locations did not return the activity they expected. Near these, the fuel may have sagged out of shape.

Next, they attached a contact microphone to a control-rod extension and listened as the liquid sodium was pumped around the primary loop. They heard a clapping sound. They slowed the pumps. The clapping sound slowed. Was there something loose in the core? It was impossible to see.

They had to feel around using the fuel-loading devices to determine the state of the core. Proceeding slowly and cautiously, it would take four months, into January 1967, to confirm that fuel had melted. First they raised the hold-down column on top of the core, and they found that it was not welded to the reactor. This was good. Next, they swung the refueling arm over the core and tried to lift the fuel assemblies, one at a time. The strain gauge on the arm would weigh each assembly. Two seemed light. Fuel had dropped out the bottom, apparently. Two were stuck together and could not be moved without breaking something. It took five months, until May 1967, to remove the fuel using the automatic equipment. Finally, seeing the fuel assemblies in the light of day, it was clear that two had melted and one had warped. There was still no explanation as to why.

The sodium was drained out of the reactor tank, although there was no provision in the design for doing so. A periscope, 40 feet long with a quartz light attached, was specially built to be lowered into the darkness from the top of the reactor vessel. Finally, the engineers could see the bottom of the reactor tank. It looked clean and neat. There was no melted uranium dripped onto the spreader cone. No loose fuel slugs were scattered around. All of the damaged fuel had collected on the support plate for the bottom of the reactor core. There was, however, something out of place. It looked like… a stepped-on beer can, lying on the floor of the reactor tank? That could explain the core melt and the clapping noise. Caught in the maelstrom of coolant forcing its way through the bottom of the reactor core, this piece of metal had slapped up against an inlet nozzle and blocked sodium from flowing past a couple of red-hot fuel assemblies. But what was it, and how did it get in the reactor?

To answer that question, the metal thing would have to be removed from the reactor, and that was not easy. Cisler stood firm under a hail of abuse from anti-nuclear factions as the Fermi 1 engineers accomplished the impossible. They built a special remote-manipulating tool and lowered it down into the floor of the reactor tank through the sodium inlet pipe. It had to make two 90-degree turns to get there. With the new tool in place, they were able to move the metal thing closer to the periscope, flip it over, and take pictures.

Still, the experts could not tell what it was. They all agreed on one thing: it was not a component for the reactor as shown on the design prints. It would have to be removed. With enormous effort and a great deal of money, another special remote operator was built and inserted through the sodium inlet. One worker swung the quartz light on the end of the periscope to bat the object into the grabber at the end of the new tool. Another man, working 30 feet away, manipulated the grabber tool blindly from instructions over an intercom. Finally, a year and a half after the accident, on a Friday night at 6:10 P. M., the mystery object fell into the grip of the tool. Slowly, taking 90 minutes, they snaked it up through the pipe and into the hands of the awaiting engineers.

They looked at it, turned it over, looked again, and finally truth dawned. It was a piece of the zirconium cover that they had attached to the stainless steel spreader cone, nine years ago. They had not bothered to have it approved and put on the final prints. It had cost an additional $12 million to figure this out.

By May 1970, all repairs had been made and Fermi 1 was ready for a restart. AEC inspectors were on hand, making the operators nervous at the close monitoring of their every action. Things were tense as 200 pounds of sodium suddenly broke loose in the primary transfer tank room, tearing out a water-pipe run and causing a loud thud as the mixture of water, air, and sodium exploded. This embarrassing incident cost another two months of down­time to repair the damage, but in October the plant finally reached its designed power level, making 200 megawatts of heat. For the next year of operation, the plant was able to remain

online for only 3.4% of the time.151 Denied an extension to its operating license in August 1972, its operation ended on September 22, 1972. The plant was officially decommissioned on December 31, 1975, the fuel and the sodium were removed, and it still sits quietly at Lagoona Beach, next to Fermi 2, a General Electric boiling-water reactor that is currently making power for DTE Energy. All things considered, Fermi 1 failed at its mission, to spearhead the age of commercial plutonium breeding in the United States. Admiral Rickover had summed it up clearly back in ’57 in one sentence, saying that sodium-cooled reactors were “expensive to build, complex to operate, susceptible to prolonged shutdown as a result of even minor malfunctions, and difficult and time-consuming to repair.”

Did we almost lose Detroit? No. There was no water in the reactor vessel to destructively explode into steam with a suddenly overheated reactor core. No steam meant that there was no source of force to break open the containment dome and spread fission products from the

core.152 The water and steam were out in another building, and the worst that could happen was to mix them with the secondary sodium loop and not with the primary loop containing radioactive sodium. The results of a massive breakdown in the secondary loop could dissolve every aluminum drink can within five miles with vaporized sodium hydroxide, but it would spread no radioactivity. The reactor was too feeble to build up enough fission product to justify the thousands of casualties predicted if the core were to somehow explode. In a maximum accident, the entire core would overheat and melt into the bottom of the reactor vessel, but it would melt the controls and the non-fissile core structure along with it, making it unable to maintain a critical mass. The dire predictions and warnings had been dramatic, but hardly realistic.

The school of business at Northern Michigan University was renamed the Walker L. Cisler College of Business. He died in 1994 at the age of 97.

Not even slightly discouraged, the AEC proceeded to secure funding for yet another stab at a commercial sodium-cooled contraption in 1970, the Clinch River Breeder Reactor Project, to be built inside the city limits of Oak Ridge, Tennessee. This would be a full-sized power reactor, making a billion watts of heat, turning out 350 megawatts of electricity, and producing more plutonium than it burned. Lessons learned from Fermi 1 were applied to the design, including multiple coolant intake passages per fuel assembly to make inlet blockages “impossible.”

Estimated cost of the plant was $400 million, with $256 million to be paid by private industry. Being built right in the middle of the Great Atomic Downturn in the mid-seventies, the project spun out of budgetary control and was plagued with contracting abuse charges, including bribery and fraud. By the time the Senate drove a stake through its heart in 1983, $8 billion had washed away, and plans for a commercial breeder economy in the United States went with lt.152

The U. S. was not alone in an early quest for a sodium-cooled plutonium breeder. In 1964, the Soviet Union under Minister of Atomic Energy Yefim P Slavsky began construction of what would become the world’s first and only nuclear-heated desalination unit making more fuel than it used, the BN-350 power station. The site was two miles in from the shore of the Caspian Sea on the Mangyshlak Peninsula. The reactor was designed to run at 750 megawatts, driving five sodium loops at the same time with one spare. It was first started up in 1972, and although it was not able to make its designed power level, it ran for 26 years. For 22 of those years, it actually had an operating license, and the fact that it kept going for so long speaks well of the operating staff. They were an exceptionally tough bunch, eventually developing an immunity to frequent sodium fires and explosions.

Startup tests immediately found weld problems in the steam generators. Both the endcaps and the tubesheets would tend to break and mix water with sodium. In the first three years of operation, there were eight sodium explosions. For some odd reason, loop No. 4 never came apart. In 1974, after two major leaks and three small ones, it was decided to rebuild all the steam generators except No. 4. By February 1975, three loops had been fixed, and they decided to start up. Seven days later, loop No. 5 disintegrated. The steam generator in loop No. 5 was replaced with one made in Czechoslovakia.

As of December 16, 1991, the BN-350 was no longer in the Soviet Union. It was in the newly formed Republic of Kazakhstan, and it was re-named the Aktau Nuclear Power Plant. It continued to supply 120,000 cubic meters of fresh water for the city of Aktau way past its projected lifetime of 1993. BN-350 fissioned its last nucleus in 1999, due to a lack of funds to

buy more fuel.154

Before the BN-350 was started up, the Soviet government was building an even larger sodium-cooled fast breeder reactor, the BN-600 in Zarechny, Sverdlovsk Oblast, Russia. It successfully started up in 1980, but a larger Russian sodium-cooled reactor means larger sodium explosions. As of 1997, there had been 27 sodium leaks, 14 of which caused serious fires, and the largest accident released over a ton of sodium. Fortunately, each steam generator is in its own blast-wall-protected cubicle, and any one can be reconstructed by the on-site workers while the reactor is running. That is one way to solve the problem of bad welds. A BN-800 and a BN-1600, still larger breeder reactors of the same type, are currently under construction in Russia.

We move on to France, a country firmly committed to a nuclear economy.

France saw the same writing on the wall that warned others about the limited supply of uranium, and in 1962 started construction of a modest-sized, experimental sodium-cooled fast breeder reactor named Rapsodie, in Cardache. It began operations on January 28, 1967, making 20 megawatts of heat. It ran okay, and in 1970 the core was redesigned and the power was increased to 40 megawatts. Under the stress of the higher output, the reactor vessel developed cracks, and it was reduced to 24 megawatts. By April 1983, the sodium leakage in Rapsodie was too costly to fix, and it was put to sleep. On March 31, 1994, a highly experienced, specialized, 59-year-old CEA engineer was killed in an explosion while cleaning the sodium out of a tank at Rapsodie. Four people were injured.

In February 1968, after Rapsodie had run for a year, ground was broken in Marcoule for a bigger, 563-megawatt sodium-cooled breeder named Phenix. Ownership and construction costs were shared by the French Atomic Energy Commission (CEA) and Electricite de France (EDF), the government-owned electrical utility. It was big and ambitious. The starting core was 22,351 rods of pure plutonium, with 17,360 depleted uranium rods in the breeding blanket and an awesome 285,789 around the periphery for neutron reflection back into the blanket and radiation shielding. Its 250 megawatts of electrical power were connected to the grid on December 13, 1973, two months after the Organization of Petroleum Exporting Countries (OPEC) halted oil deliveries to countries that supported Israel and increased the market price of crude oil by a factor of four. The first sodium leak occurred in September 1974. In March and July there were two more, causing slow spontaneous combustion in pipe insulation.

The engineering innovation of Phenix was a free-standing or “free-flowering” core restraint, allowing it to expand or contract as it wished due to thermal, mechanical, or irradiation effects without bending or breaking anything. It ran perfectly for 16 years, as the spot price for uranium went from $6 per pound in 1973 to $40 per pound in 1976. Running on plutonium seemed a splendid idea. On July 11, 1976, a sodium fire broke out at the intermediate heat exchanger. Another one on October 5, and three more by the end of 1988. In May 1979, the fuel cladding failed, releasing radioactive xenon-135.

On August 6, 1989, something very odd happened to Phenix. A sudden very negative reactivity excursion triggered a scram. The reactor simply quit making neutrons, and the power level fell like a lead brick. The engineers had no idea why. The reactor restarted without a problem, but 18 days later it happened again. This time, the instruments were blamed for both negative excursions, but nothing wrong could be found. The ability to make electrical power

dropped to near zero as the incidents continued.155

On September 14, 1989, the power again went to zero. The cause was believed to be a gas bubble in the core periphery. The problem was solved with mechanical maintenance, and Phenix was restarted in December. It happened again on September 9, 1990, and the gas — bubble hypothesis went out the window.

For the next 12 years, panels of experts pondered the problem, and the plant was tested, taken apart, put back together, repaired, modified, and refurbished, not generating any power to speak of. No one specific scenario or cause of the strange incidents was identified, but the only thing that made sense, given the speed and extent of the events, was that the core was in motion, moving in different directions as it generated power and disturbing the critical configuration of the reactor as designed. Lesson learned: avoid free-flowering core restraints in future reactor designs. In June 2003, Phenix was restarted and ran at reduced power, 130 megawatts, until final shutdown in 2009.

In 1974, a European fast-neutron reactor consortium, NERSA, was established by France, Germany, and Italy to build the biggest plutonium-fueled breeder reactor in the world. This extraordinary allegiance of countries lasted about a year. Germany spun off and decided to make their own breeder, the SNR-300 in Kalkar, and in the middle of 1976 President Valery Giscard d’Estaing of France proclaimed that his country would build the Superphenix, a 3- gigawatt sodium-cooled breeder, at Creys-Malville, 45 kilometers east of Lyon. Italy stood still and did not make a sound.

Minister of Industry Andre Giraud announced to a spellbound crowd at the American Nuclear Society meeting in Washington, D. C., that there would be 540 Superphenix-sized breeder reactors in the world by the year 2000, and 20 of them would be in France. Meanwhile, 20,000 people occupied the building site to protest the thought of another reactor project. On July 31, 1977, the protest got serious, with 50,000 participants, and riot police went in armed with grenades. A local teacher, Vital Michalon, was killed, another protester lost a foot, and a third lost a hand in the battle. Ongoing protest and sabotage made construction work difficult, and on the night of January 18, 1982, militant Swiss eco-pacifists fired five rocket-propelled grenades

at the containment building, causing cosmetic damage.156

Construction continued under bombardment, and the completed Superphenix went critical on September 7, 1985. It was connected to the grid on January 14, 1986. A series of administrative hurdles and incidents prevented any significant power production until March 8, 1987, when a massive liquid-sodium leak was discovered issuing forth from the refueling rotor tank (storage carousel). The reactor was down until April 1989 as an alternate refueling scheme was designed. The rotor tank was too far down in the guts of the reactor, and there was no way to repair the leak, a crack 24 inches long, without dismantling the entire plant.

Operation at very low power proceeded until July 1990, when a defective compressor was found to be blowing air into the liquid sodium line and making solid sodium oxide. In December the roof of the turbine hall collapsed in a heavy snowstorm. Superphenix seemed cursed.

The reactor was shut down for good by government decree on December 30, 1998. It had remained offline for two years because of technical difficulties and four and a half years for administrative debating. It never ran at its designed power level. Superphenix is scheduled to be completely dismantled by 2025. No new breeder reactor is planned in Europe.

The Germans’ 327-megawatt SNR-300 sodium-cooled fast breeder reactor, costing a scant $4 billion to build, was completed, filled with sodium, and ready to be started up in 1985. Political hand-wringing kept it in standby condition for six years, costing about $6 million per month to keep the sodium liquefied using electric heaters. On March 21, 1991, the project was officially cancelled. The SNR-300 and the ground it was sitting on were sold at auction for half the monthly upkeep cost to a Dutch developer, who turned it into an amusement park named

Kernwasser Wunderland.

Of all the countries building sodium-cooled fast breeder reactors, India would be voted least likely to pull it off without blowing something to kingdom come. A list of sodium fires, explosions, and inexplicable power excursions in their Fast Breeder Test Reactor (FBTR) would be monotonous, but one accident stands out as unique. The reactor was first started up in October 1985. FBTR is a copy of the French Rapsodie, built at Kalpakkam using elephants for heavy lifting. Like every other fast breeder made since 1957, the blind fuel-handling machinery was based on the ingenious and complex Fermi 1 design.

In May 1987, a fuel assembly was being transferred from the reactor core to the radiation shield at the periphery. Each assembly held 217 fuel rods in a square metal matrix. For reasons unknown, one fuel assembly was sticking up one foot above the core as the handling arm swept across to find the one it was supposed to pick up. The electrical interlock that normally prevents the arm from moving if anything is in the way had been bypassed. Crunch. The arm rammed the protruding assembly, bending it out of shape, and then knocked the heads off 28 assemblies in the reflector as it tried to back away. Trying to make everything back like it was, the arm mechanism accidentally ejected an assembly in the reflector and put a 12.6-inch bend in a substantial guide tube. It took two years to sort out the damage and repair the core. Reasons for the accident were never really understood.

India is presently building two larger sodium-cooled fast breeder reactors. Be afraid.

The Japan Atomic Energy Commission published its first Long Term Plan in 1956, and at the top of the list of technologies to be developed were a sodium-cooled fast breeder and an associated fuel cycle using plutonium extraction. Of all countries in the game, Japan had the strongest incentive to build breeder reactors and become energy-independent. Domestic power options were few, and they had recently fought a war over energy resources and lost.

Two breeder reactor projects were started in parallel. The smaller unit, Joyo (Eternal Sun), achieved criticality on April 24, 1977, developing 50 megawatts of heat energy. Located in Onari, Ibaraki, this reactor is now on its third core-loading, and the power has been increased to 140 megawatts. It has been used as a test-bed for fuel mixtures and materials for use in future breeder reactors.

The second breeder, Monju, was built in Tsuruga, Fukui Prefecture, and was first brought to

criticality in April 1994.157 It is less of a test reactor than Joyo, producing 280 megawatts of electricity from 714 megawatts of heat. It has three double-loop core coolers, A, B, and C. The inside loop is full of radioactive sodium, and the outside loop, of course, has non-radioactive sodium, connected to the steam generator.

All was well with Monju until December 8, 1995. It was operating at 43% power when a smoke alarm went off near the hot-leg pipe for outside loop C, about where it exited the reactor vessel. It was 7:47 P. M. High-temperature alarms went off. The operating crew started a slow controlled shutdown 13 minutes later. It was beginning to look like a sodium leak, so at 9:20 P. M. they went ahead and hit the scram button. By 12:15 a. m. they had successfully drained the outside C-loop of sodium.

Turbulent flow in the sodium pipe had caused intense vibration, which broke off a sealed thermocouple well inside the pipe and bent the thermocouple 45°. Hot, liquefied sodium oozed through the now-opened thermocouple penetration in the pipe to the electrical connection and started dripping on the floor. Eventually, three tons of sodium collected in a clump on the concrete, and it caught fire, hot enough to warp and melt steel structures in the room. No one was hurt, and no radioactivity was released, but there was a lot of eternal shame.

The government-established Power Reactor and Nuclear Fuel Development Corporation (PNC) tried to cover up the accident, but when accounts broke free there was a public and political uproar and questions as to what else they had not divulged. A restart after repairs was delayed until May 8, 2010. Fuel was replaced, but on August 26, 2010, a 3.6-ton in-vessel transfer machine being hoisted into place slipped its bindings and did a free dive into the reactor vessel. It was somewhat mangled and could not be retrieved from the opening through which it had fallen. A lot of engineering work by the Japan Atomic Energy Agency achieved removal on June 23, 2011. To this date, Monja has managed to generate power and put it on the grid for one hour. Another, larger fast breeder in Japan is planned, but plans have gone awry in recent years.

Given this interesting set of mini-disasters in which there were no injuries, one would have to consider the truth in Admiral Rickover’s terse assessment of liquid-metal-cooled reactors. Take a step back for the longer gaze, and it starts to look like the most awful way to build a machine that has ever been designed. The people who championed and worked hard for this new and dangerous technology were visionaries. There was actually no particular need for an advanced power system that would last forever and free entire nations from a dependence on others. In the 1950s, there was enough burnable material, oil, coal, and gas, to go around, and enormous reserves of uranium were discovered throughout the decade. These individuals were forcing us into the future with all the speed we could handle, developing novel and outrageous concepts, materials, and machinery, to a place where mankind would eventually have no choice but to go. They were just too early. Was this wrong?

Next, we will explore an area in which the man-machine interface is pushed to the limit, and people die.158

134The USS S-48, built in 1921, was possibly the only submarine to find peril in a heavy snowstorm. She was returning home to Portsmouth, Connecticut, on the night of January 29, 1925. At about 6:30 P. M. off the coast of New Hampshire, the wind picked up and heavy snow started falling. With visibility zeroed, she ran onto the rocks off Jeffery Point, rolled off, then grounded again in Little Harbor. Early next morning, still stuck and listing badly the battery compartment started taking water, and the ensuing electrolysis action broke down the salt water and released chlorine gas inside the sub. The Coast Guard picked up the crew an hour and a half later, and they were transferred to the hospital at Fort Stark for treatment.

135At least no leakage problem while running. The USS Seawolfsubmarine with a liquid-sodium-cooled fast reactor did experience an unscheduled sodium expulsion, but the boat was in dock at the time. The Superphoenix fast breeder reactor in Creys-Malville, France, has suffered from sodium leaking and corrosion in the cooling system, and it has been out of service since September 1998.

136Otto K. Helney a 53-year-old engineer, Larry A. Pugh, a 51-year-old physicist, and Lee Wells, a 62-year-old assistant, seemed old enough to know better when they mixed 10 pounds of gycidal azide polymer and nitrocellulose (gun cotton) together in an aluminum pan, hoping to measure the shock wave when the two chemicals ignited. Helney and Pugh died instantly and Wells was blown against vertical terrain and burned over 20% of his body Six other men died at Santa Susana in various incidents in the early 1960s. It was an exciting place to work.

137Zirconium was an excellent choice. It is transparent to neutrons and able to withstand very high temperatures, but it was expensive. For graphite moderator units out on the periphery of the reactor core where the temperature would never be very high, stainless steel was substituted for the zirconium.

138From whose stockpile would this U-233 come? U-233 does not occur in nature, and it must be made by production reactor from Th-232. Experimental U-233 reactor fuel probably came from a stockpile at Los Alamos. In 1955 a composite Pu-239/U-233 atomic bomb named MET was built and tested in Operation Teapot. The U-233 was made using the B Reactor at the Hanford site, and this material was used in several thermal breeder experiments in the late 1950s and early 1960s. The stockpile wound up at Oak Ridge, and a $511M contract from the DOE to dispose of this material is currently under way.

139Electromagnetic sodium pumps were not unknown to the SRE design group. In fact, one was used to load the cold trap, designed to remove oxidized sodium and any other contaminants from the primary cooling loop. This pump failed about 60 days into operation at power in late June 1957. Similar electromagnetic devices, called “eddy current brakes,” were used on the primary and secondary coolant loops to stop the coolant from moving after a reactor scram.

140I can’t tell exactly where the fire occurred. It was in the “XX vault,” where X indicates an unreadable character in the SRE operations log summary There was an NAS vault, a main pipe vault, an aux pipe vault, and a primary sodium fill tank vault. There was no tetralin in any of these locations.

141 The window was three feet thick, but not the glass. Hot cell windows were made of two very thick panes of glass, mounted in parallel and spaced about three feet apart in a sealed tank, very much like an aquarium. The space between the glass windows is filled with zinc bromide dissolved in water. Objects in the hot cell are given a yellow cast by looking through the zinc solution, which is an excellent high-energy gamma-ray shield.

142The “nack” cooled seal used a mixture of sodium and potassium, or NaK, as the cooling fluid into the sodium seal to keep it frozen solid. Nack is liquid at room temperature, and it flows as freely as water. It was used in the fuel pins to fill the space between the solid uranium fuel pellets and the stainless steel tubes in which they were encased, giving a thermal coupling that would ensure heat flow to the coolant, which flowed around the tubes. The tetralin was a low-viscosity solvent, and it was hard to keep it from finding a way past the sodium seal and into the coolant. The nack would leak less, and if it did, so what? In all the other motorized sodium pumps in the system, tetralin was retained as a coolant.

143At least I assume it was a scram actuation. In the SRE plans, this switch is listed as a “setback ” actuator. Does this indicate that it was meant only to limit the power rise rate and not stop it completely? If so, it would not have worked as planned. With power increasing exponentially, a scram is quite appropriate.

144The Hallum Nuclear Generating Station is a little-known, second graphite-sodium reactor built near Lincoln, Nebraska, starting in 1958. This experiment was also funded by the AEC. Unlike the SRE, the vertical-shaft centrifugal sodium pumps for this reactor were designed from scratch and were considered far superior to the ill-fated SRE pumps. Unfortunately, the modified Hallum-type pumps used in the rebuilt SRE had undersized overflow loops which produced a serious gas-entrapment problem. The Hallum reactor operated only briefly from 1962 to 1964. The graphite moderator cans were clad in stainless steel, and stress cracking and corrosion caused irreducible problems. By 1969, evidence of the Hallum reactor was erased from the prairie, but the Hallum-type pump remains as a credible means of moving liquid sodium.

145Walter worked as a “wage slave” at the Ford Motor Company starting in 1927. Henry Ford sent him to Nizhny Novgorod, Soviet Union to help build a tractor factory but he became overly interested in the proletarian industrial democracy and Ford fired him in 1932. After working for a few years at an auto plant in Gorky Reuther returned to the U. S. and became a very active member of the UAW. On May 26, 1937, at 2:00 P. M., he and Richard Frankensteen were in the middle of a leaflet campaign (“Unionism, Not Fordism”) and they were asked by a news photographer to pose on the pedestrian overpass in front of the Ford sign. Ford’s modest army of about 40 security specialists walked into the frame from the left and proceeded to discipline the uninvited visitors. The Dearborn police stood out of range and shouted advice while the union men were beaten, kicked, dragged by the feet, slammed down on the concrete, and thrown down two flights of steps. Reuther, always a champion of the underpaid and the underappreciated worker with too little money died on May 9, 1970, when his privately chartered LearJet smacked the runway at the Pellston, Michigan, airstrip in rain and heavy fog. He was on his way to the UAW recreational facility at Black Lake.

146A breeder reactor is supposed to run off fuel that it produces, which in this case was plutonium-239. Until it has actually made fuel and it has been extracted from the breeding blanket, the breeder usually starts off with a load of uranium-235. Unlike a water-cooled reactor, a fast breeder runs off pure fissile isotope, so the starter fuel must be highly enriched to simulate its eventual load of Pu-239.

147When I was in school, fast breeder reactors were all the rage, and a wild story about the refueling machinery in the Fermi 1 was going around. It was said that a spin-off of the breeder reactor project was the AMF pin-spotting machine, now used in many bowling alleys. The method used for collecting and placing fuel assemblies in an automatic, utterly blind situation under liquid sodium was found useful for doing a similar job with bowling pins. I am sad to say that I can find no connection between bowling alley machinery and fast reactor refueling devices, much as I would like to. The patent for the AMF “candlepin” pin spotting machine was filed on April 14, 1964, which is close to when the Fermi 1 was designed, but it was based on the original AMF 82-10 pin spotter, first demonstrated in 1946.

148An enduring question of fast reactor design is: “What is the sodium void coefficient?” If the sodium coolant leaks out of a fast reactor, does it go subcritical or supercritical? The correct answer if this question should come up on your Ph. D. qualifying exam is: “That depends.” There are many factors that come to bear, including the geometrical shape of the core.

149It is difficult to find a cross-section view of the Fermi 1 reactor that does not have a big X drawn through the refueling car. It was not a popular accessory.

150Geek-joke from 1975: How does a Digital Equipment Corporation computer-repair technician fix a flat tire? He swaps the wheels to see if the problem moves. You had to be there… .

151This number, 3.4%, is the capacity factor for Fermi 1. In 2011 the average capacity factor for a nuclear power plant in the United States was 89%. A wind-turbine plant, such as the Burton Wold Wind Farm, consisting of ten Enercon E70-E4 wind turbines, had a capacity factor of 25% in 2008. The Hoover hydroelectric dam has an average capacity factor of 23%. The capacity factor of a power plant is the amount of energy produced over a set period of time divided by the energy it could have produced if working at full capacity during the same period.

152No steam not only meant that there could be no steam explosion, it also meant that there could be no explosive hydrogen buildup. Hydrogen gas can be made from steam in a reactor accident by heat-induced corrosion of metals or radiolysis. More on this topic in the Fukushima chapter.

153Meanwhile, in 1975 the Shippingport Atomic Power Station in Pennsylvania, a pressurized water reactor based on a Westinghouse aircraft carrier engine, was converted into a thermal breeder using a thorium-uranium-233 fuel cycle. The Shippingport reactor had been running since May 26, 1958, and was the first commercial reactor making significant electrical power in the U. S. The reactor ran smooth and quiet for 25 years, and it never gave any trouble.

154Even a breeder reactor has to buy fuel. If the supply services had been operating as planned, the plant would exchange its spent fuel and breeding material for a fresh load of plutonium and depleted uranium. Plutonium would be extracted chemically from the spent load, and there would be enough extra in the breeding stock to more than pay for the chemical separation and fuel-rod fabrication. The BN-350 started out with an expensive load of 20% enriched uranium with some mixed uranium-plutonium surplus from the Soviet nuclear weapons being demilitarized. By 1999, the now-obsolete fuel-rods would have to be custom-built.

155Subsequent investigations found that similar incidents had occurred in April 1976 and June 1978. They were explained away as “control rod slippage,” which turned out to be wrong.

156The unnamed group of terrorists obtained a Soviet RPG-7 shoulder-fired rocket launcher and eight rockets (“bonbons”) from the German “Red Army Faction” via the Belgian counterpart Cellules Communistes Combattantes. They lost three of the missiles in the dark. Chaim Nissim, elected to the Swiss Geneva cantonal government for the Green Party in 1985, admitted 22 years later to leading the attack. He should stay out of France.

157Monju is the Japanese word for Manjusri, a bodhisattva or enlightenment-being inMahayana Buddhism. The Japanese tradition holds that Monju invented male homosexual love. What this has to do with a sodium-cooled fast breeder reactor, I’m not sure.

158I had to leave out the other participant in the sodium-cooled circus, Great Britain, not because they never had a coolant leak, but because the British incidents were too routine to stir the reader. There were two notable British fast reactors built at Dounreay near the chilly northernmost frontier of Scotland. There was the Dounreay Fast Reactor (DFR), which was nack-cooled, and the Prototype Fast Reactor (PFR), cooled with sodium. DFR was started up on November 14, 1959, and PFR was the last to shut down permanently in 1994. The U. K. push for a plutonium economy has since been declared dead.

Chapter 7