If the power got out of hand for some unforeseen reason, the reactor would scram automatically on the gamma-ray level signal at a pre-set point, below the level where any harm could come to the machinery. As a backup, a set of thermocouples in the core monitored the reactor temperature continuously, feeding another scram channel. There was no steam to explode, so the worst thing that could happen would be too much power making too much temperature and causing the thing to melt.
The power level passed the 60-kilowatt level. No problems. They were now in untested territory, but the engines were accelerating smoothly. 80 kilowatts. No problem. 90 kilowatts. Still increasing power. 96 kilowatts. The power level started to fall rapidly, as if the bottom had dropped out. The crew, watching the instruments, found this curious, and the reactor operators cocked forward in their seats. The automatic control, sensing that the power was dropping quickly, pulled out the controls, trying to bring it back up. The power seemed to keep going south, and the needle on the meter fell to the left. Twenty agonizing seconds passed and WHAP! The scram circuit, detecting that the thermocouples in the reactor core had all melted, shut her down automatically, throwing in all the controls at once. In those 20 seconds, every piece of fuel in the reactor had lapsed into the liquid state. The fact that it overheated had not disassembled the reactor, as it was made of some very rugged ceramic materials, and the reactivity had actually increased as the nickel-chromium-uranium-oxide fuel turned to fluid. When the indication had been that the power was falling, it was actually increasing very quickly. Although the fuel was designed to be very tolerant of extreme temperature, the power level had spiked beyond the capacity of the two jet engine air-intakes to keep the reactor core from melting. Only a few of the zirconium-hydride moderator sections were damaged.
A slight increase in background radioactivity was detected downwind of the smokestack as some few fission products, evaporated off the white-hot fuel, made it through the filter bank. Aside from that and the pen-chart recording, there was no outside indication that anything had gone wrong, and no humans were harmed.
The steel annulus surrounding the reactor was pumped full of mercury from a holding tank. Mercury is a high-density metal, liquid at room temperature, and it makes an excellent gamma-ray shield. With it in place, the crew could approach the Heater-Three to disconnect it without danger from the decaying fission products built up in the damaged fuel.[103] The engine was then dragged back to the hot-lab in the assembly building, where the core was rebuilt and the source of the problem was analyzed.
That filter in the high-voltage cable had unfortunately limited the number of electrons per second that could travel through the wire. That was fine, as long as not too many electrons were needed to register the number of gamma rays per second that were traversing the ion chamber. At the higher power level, at which the equipment had never been tested, the gamma flux overwhelmed the ability of the power supply to keep up. The current demand from the chamber was so high, the voltage dropped, and the detector stopped detecting. The automatic system interpreted this as a power loss, and it tried to compensate for it by pulling the controls. The power climb accelerated until the engine was, as we say, outside its operating envelope. This incident went down as the first time in history a reactor was melted because of an instrumentation error. Human error was a fault only indirectly.
In December the AEC told the Air Force that the nuclear bomber could not be flown over the United States. The only way it could be flown was out over the Pacific Ocean, presumably taking off from the beach. The 8-million-dollar shielded hangar, recently finished, could no longer be used to house a nuclear bomber, and grading the runway would not be productive. On January 20, 1961, John F. Kennedy was sworn in as President of the United States. On March 28, he signed a paper canceling the ANP project, and that was that. The much disappointed staff at the NRTS knew that they were very close to a working atomic aircraft engine, but for the good fortune of nuclear power we will never know if it would have flown.
A cluster of three reactor accidents involved the explosive conversion of water into steam by nuclear means. Two of the incidents were accidents only by procedural technicalities, and the other one was a complete surprise. We know exactly how it happened, down to the millisecond, but we have no answer as to why.
Among the gifted reactor physicists at Zinn’s Argonne lab was Samuel Untermyer II, grandchild of Samuel Untermyer I. Sam the First was a Jewish-American born in Lynchburg, Virginia, who became the most famous New York lawyer of all time. He was responsible for, among other things, maintenance of the 5-cent subway fare. His grandson, a graduate from MIT in 1934, did not believe that the revered pressurized water reactor, in which the coolant is kept at such a high pressure it cannot boil, was the only way to build a water-moderated pile. In those early days of reactor development in the late 1940s, the general opinion was that if the water in the reactor vessel were allowed to boil, then the neutron production would become erratic and unpredictable. The coolant voids caused by steam bubble formation were predicted to cause fuel melts, chemical explosions, unchecked power excursions, and probably boils on the reactor operators and tension headaches.
Untermyer proposed a contrary prediction. If the coolant, which is also the moderator, in the reactor vessel were allowed to boil into steam, it would simplify the power production mechanism. There would be no need for a complicated, expensive, failure-prone steam generator and a second cooling loop. The reactor vessel would become a boiler, like the boiler in a coal-fired power plant but simpler, without any boiler tubes. No pressurizer to maintain the lack of boiling would be necessary, and several pumps and valves could be eliminated. Instead of running wild and unpredictable, the neutron flux would be controlled. A great deal of boiling would result in moderator voids, which would degrade the fission process and lower the power level. If the boiling were quenched, the density of the moderator would increase, and the power level would rise to a threshold cutoff, floating between too much power and too little power. Such a boiling-water pile would control itself and follow the load demand. There would be a lesser need for an electronic feedback-control system, such as would cause the HTRE-3 to melt in ’58. In the event of an uncontrolled upward power excursion, the moderator would boil away, and this would automatically turn off the fission process without the need for an electronically controlled scram system. These seemed like desirable qualities in a power reactor, but it would all have to be proven with physical experiments.
Untermyer convinced his boss, Walter Zinn, of the importance of the boiling-water-reactor concept, and together they petitioned the AEC for a contract to prove the principles. In 1952 he was given enough money to make a modest stab at it and a spot of desert at the NRTS. The test reactor would be laid out on the ground, without so much as a tin roof over it, and the control room was a small trailer parked half a mile away. Television cameras gave the experimenters views of the reactor, including one using a large mirror to show the top of the core. The reactor vessel was a steel tank, half an inch thick, 13 feet high and 4 feet in diameter, halfway sunk in the ground. The 28 aluminum-clad enriched-uranium fuel elements were surplus from the big-budget Materials Test Reactor being erected elsewhere on the desert. The total fuel loading was 4.16 kilograms of uranium. The control rods were inserted through the open top of the vessel, adjusted in and out of the core with electric motors and wired back to the control trailer. Pipes, cables, and tanks were all over the ground. The steam made by the reactor heat was simply vented into the air. Dirt was mounded around the exposed portion of the reactor to give it some gamma-ray shielding. Untermyer, expecting this to be the first in a glorious series of experimental reactors, named it BORAX–I.[104]
103
Hold on to your seat. Three quarters of the free world’s supply of mercury was used to shield the HTRE-3. At the end of the project it had to be released slowly back into the market to prevent an economic crash in the liquid metals trade.
104
Although I and everyone else call it BORAX–I, it may not have been known as a numbered unit until BORAX–II was built. Contemporary reports from the field simply call it “The Borax Experiment.” Local newsmen called it “The Runaway,” and this name stuck to it like cobalt-60 contamination.