This article focuses on learning from the successes and failures of the first-generation reactor development. Reactor designs have evolved over time to meet increasingly rigorous demands for safety and to take advantage of technological developments to improve their economics, but these changes have been piecemeal. Although light-water reactors are the most common reactor technology in use today, heavy-water reactors were actually developed earlier. The earliest demonstration of a heavy-water moderated and cooled reactor took place in May 1944 at Argonne. The reasons for the domination of water-cooled reactors, and particularly of light-water reactors, are complex. The article suggests that it is interesting to speculate on how the new initiative to develop more advanced designs may play out. There are already strong pressures to focus on the integral light-water design; based on well-understood light-water technology, the argument goes, such designs will be much easier to develop and license. In the longer term, however, some of the non-light water reactors could ultimately achieve greater levels of passive safety, efficient fuel utilization, economic performance, and proliferation resistance.
There are over 400 commercial nuclear power reactors in operation around the world today. Almost all of them use enriched uranium, and all but a handful are moderated and cooled by water. Reactor designs have evolved over time to meet increasingly rigorous demands for safety and to take advantage of technological developments to improve their economics, but these changes have been piecemeal. Therefore, in the year 2000, when I was the deputy director of the U.S. Department of Energy's Office of Nuclear Energy, we convened a meeting of our counterparts from a small number of nations with active nuclear research programs to discuss collaborating on the development of the next generation of reactors. We hoped that these new reactors would result in improvements in reactor safety, economics, proliferation resistance, fuel use, and the generation of waste.
The consensus was that, although the current fleet of operating reactors was safe and economical, various pressures were leading to a need for more advanced designs that required substantial departures from the existing technology. These advanced concepts were considered revolutionary rather than evolutionary, and were dubbed “Generation IV” reactors—Gen IV, for short.
After collectively reviewing over 100 variations of different advanced reactor concepts, the group narrowed down the list to six categories of advanced reactors. (That subject was covered in “Nuclear's Next Generation” in ME Power & Energy, June 2003.) As the members of the Gen IV group began to discuss the six concepts within the nuclear communities in their respective countries, an interesting observation began to emerge: Most of these new concepts were not new at all. Older scientists and engineers would raise their hands at presentations about the Gen IV reactor program and gently remind the speakers that work had been conducted on similar designs as far back as the 1940s and 1950s. In fact, some of those raising their hands had been the very scientists and engineers who had done that earlier work.
These advanced concepts were considered revolutionary … but work had been conducted on similar designs as far back as the 1940s.
The new generation of these designs is not identical to the old. At the very least, the current generation of designs can take advantage of advances in materials and technological understanding that did not exist half a century ago. But it is true that some version of all of these generic concepts were among the early designs studied.
In my capacity as deputy director, I gave many briefings on the program, and I was challenged often on that very point. As a result, I became interested in some of the well-known and lesser-known aspects of this early history of nuclear power and began to research them in my spare time. I found it intriguing to learn how many years ago we’d actually begun to develop some of the technologies that are today regarded as “advanced,” only to abandon them along the way. I also began to discover some aspects of the story that were not so well known—and in fact, in some cases, that contradicted what I always thought to be true. I collected these into a book, Nuclear Firsts, published in 2010.
The lessons of these nuclear “firsts” may be instructive as we look toward developing the next generation of reactors.
Two of the Gen IV concepts—the Very High Temperature Reactor and the Gas-cooled Fast Reactor—are cooled by gas, not water. Gas-cooled designs are used in the United Kingdom but nowhere else at the moment. But gas-cooled reactors date back to the very beginning of nuclear power.
The earliest reactors, which were very small and were not intended for power generation, were not even called reactors. They were called “piles,” and they were essentially stacks of large graphite blocks with channels containing the uranium fuel. The very first of these reactors, the Chicago Pile (designated CP-1), reached initial criticality on Dec. 2, 1942, and provided the proof of principle that a sustained fission reaction could be achieved. It was essentially a zero power reactor, so only natural circulation of air through spaces between the blocks was required.
Within months, researchers at Oak Ridge, Tenn., began building a larger graphite reactor. Called the X-10 Graphite Reactor, it began operation in November 1943 and was the first reactor in the world to operate above zero power. Large fans forced air through the channels in the graphite to remove the heat generated by the chain reaction. Although the energy produced was only about 1,000 kilowatts thermal initially (and later scaled up to 4,000 kWth), the fans required were among the largest available at that time.
The X-10 Graphite Reactor was built as a pilot facility to test processes for the production of plutonium from uranium for the weapons program. The full-scale production facility would later be built at Hanford, Wash. (The Hanford reactor was a graphite reactor as well, but used water for cooling.)
After the X-10 Graphite Reactor had fulfilled its primary mission of supporting the development of the Hanford production reactors, it was used for several other pioneering activities. One of these was that, in August 1946, it became the first reactor in the world to produce an isotope (carbon 14) for peaceful purposes.
Its second “first” was, however, a real surprise to me. In September 1948, researchers at Oak Ridge hooked up a toy generator to the reactor, attached the generator to a flashlight bulb, and lit the bulb with the first trickle of electricity from a reactor—about one-third of a watt. Although all accounts credit the Experimental Breeder Reactor-I (more about that below) as producing the first electricity, the Oak Ridge reactor accomplished it first, albeit in minuscule quantities.
Two other Gen IV concepts are liquid-metal reactors. The Sodium-cooled Fast Reactor and the Lead-cooled Fast Reactor also have antecedents from the beginning of the atomic age. In August 1946, a mercury-cooled reactor at Los Alamos began operation as part of the weapons program that continued following the end of World War II. That reactor accomplished a couple of firsts: it was the first liquid-metal-cooled reactor and the first unmoderated, or “fast,” reactor to operate. The reactor was fueled with plutonium, and because the wartime code word for plutonium had been 49, the scientists and engineers at Los Alamos christened it “Clementine” after the song about the Forty-Niners of the California gold rush.
Another liquid-metal-cooled reactor was the first to produce electricity in quantities greater than a flashlight battery. The Experimental Breeder Reactor-I, built in Arco, Idaho, used a potassium-sodium alloy for removing the heat generated. Although the reactor was intended to demonstrate breeding (and it ultimately did, although in very tiny amounts), its initial use was to produce the first usable quantities of electricity. On Dec. 20, 1951, just a day after the reactor first reached full power, it was hooked up to a Rankine steam engine and used to light four 200-watt incandescent light bulbs. The following day, the reactor output was raised to 100 kilowatts and was briefly used to supply power to all the electrical equipment in the reactor building.
It is not difficult to understand why this event eclipsed the earlier demonstration of electricity from the Oak Ridge reactor. In the first place, the tight veil of secrecy that had applied to all things nuclear during, and shortly after, the war was slowly lifting. Some claim the Oak Ridge experiment was kept quiet because it hadn’t been authorized. Furthermore, the Oak Ridge demonstration was so small that it might have been difficult to see the practicality of the technology. The Arco demonstration, by contrast, although still very small, was orders of magnitude larger. And not incidentally, the results were visible in a photograph. In fact, the image of the four light bulbs has become an iconic emblem in the annals of nuclear power development.
In 1979, ASME named the reactor a National Historic Mechanical Engineering Landmark.
Even though the current global fleet of nuclear reactors is dominated by designs that rely on water for cooling and moderation, only one of the six Gen IV concepts, the Supercritical Water-Cooled Reactor, advances light-water technology. Historically, work on light-water moderated and cooled reactor designs actually began well after initial efforts with gas and liquid-metal technology. Two reactors that could be called forerunners of the present light-water reactors began operation in 1950. The Zero Power Reactor at Argonne and the Bulk Shielding Reactor at Oak Ridge were built in support of the nascent nuclear submarine program, and in fact, the submarine program is usually regarded as the impetus that propelled light-water reactor technology to the forefront when nuclear technology began to be commercialized.
That outcome was not necessarily obvious at the start of the Navy's program. The program, operated under the command of Admiral Hyman G. Rickover, was a substantial effort designed to develop and test reactors that would be optimized for operation in a marine application. The early program sought to develop both (pressurized) light-water and liquid-metal (sodium) designs, and sea trials of both technologies were conducted. Ultimately, light-water design prevailed and became the basis of the Naval nuclear program.
As a result of Rickover's influence, when the Atomic Energy Commission decided to provide a “large-scale” demonstration of the feasibility of nuclear power by building a reactor in a utility environment, the pressurized water reactor design was selected. This reactor, known as the Shippingport Atomic Power Station, began producing power on Dec. 18, 1957. At 60 MWe, Shippingport is generally credited with being the first fully civilian reactor of a significant size in the world.
However, Shippingport was not the first reactor to supply power to a commercial grid. Starting in 1954, a number of reactors were connected to grids, including a small liquid metal reactor in the Soviet Union, a larger gas-cooled reactor in the United Kingdom, and several small reactors of various types in the United States (including pressurized water, boiling water, and liquid-metal designs). Each of these other reactors moved the state of the art forward in some way and has a legitimate claim to being a “first” by one measure or another. However, they all were either significantly smaller than Shippingport or were reactors designed primarily for the production of plutonium and only supplied power to the grid secondarily. Shippingport, also an ASME Landmark, was the first reactor of a commercially viable size to operate solely for electricity production.
Although light-water reactors are the most common reactor technology in use today, heavy-water reactors were actually developed earlier. The earliest demonstration of a heavy-water moderated and cooled reactor took place in May 1944 at Argonne. Known as Chicago Pile No. 3, it was one of a series of tests of different possible fuels, structural materials, and coolants for reactor designs. Heavy-water reactors in the U.S. never progressed past that point, but a year later, when the Canadians began to explore nuclear technology, their first experiment, a zero-power reactor that began operating in September 1945, used heavy water for moderation. As they built larger designs for research, and later for power production, they also used the heavy water for cooling. All power reactors in Canada use heavy water, and the Canadian technology has been exported to other countries as well.
Of the six Gen IV concepts, only the Molten Salt Reactor has never been deployed for commercial electricity production. But a number of molten-salt research reactors were built and the concept was once regarded as very promising.
While early versions of these designs may not have succeeded in the marketplace, advocates believe the situation is different now.
Molten-salt reactors were initially developed because of the attractive features they offered for powering aircraft and even spacecraft. Molten-salt reactors could theoretically achieve high power densities, which made them ideally suited for applications where weight was a serious limitation. In fact, when the first molten-salt reactor began operation at Oak Ridge in 1954, it was called the Aircraft Reactor Experiment.
Unlike most other reactor designs, where a coolant circulates around solid fuel, in the ARE, the fuel was in liquid form and circulated around the moderator of the reactor. Several other experimental molten-salt reactors followed the Aircraft Reactor Experiment. Although the demonstration reactors met technical expectations, the program was never carried through to completion. A molten-salt reactor was loaded onto an aircraft and operated on the aircraft, but was never used to power the aircraft.
In addition to aircraft propulsion, the use of molten-salt reactors for commercial power production was also considered, but never pursued to completion. (Molten-salt reactors were the subject of “Too Good to Leave on the Shelf” in the May 2010 issue.) By this time, the focus was on light-water reactors.
Attempts to explore still other reactor technologies have so far had limited results. Those attempts included one power-producing reactor using an organic moderator and coolant operated in Piqua, Ohio, in 1963, and several pebble-bed reactors in Germany, starting with the AVR in 1967. Although there has been no further interest in organic reactors to date, the pebble-bed concept was considered seriously a few years ago. However, that effort is currently stalled in the West, although work continues in China.
The reasons for the domination of water-cooled reactors, and particularly of light-water reactors, are complex. The early support of pressurized water reactors through the Naval nuclear program in the United States was a critical factor. The program provided substantial funding that pushed the development of this technology, and resulted in significant political interest in it. The cult of personality was important, too: Admiral Rickover clearly was a forceful figure. Changes in program priorities and budget pressures, which still sound familiar today, also took a toll. And the early success of Shippingport no doubt was persuasive with utilities looking to enter the nuclear marketplace as quickly as possible.
Some of the alternative concepts lost support as a consequence of poor performance or ill-timed accidents. The U.S. operated one gas-cooled reactor, at Fort St. Vrain in Colorado; problems with one component, a helium circulator, plagued its operations and led to its early shutdown. Several countries, including the U.S. in the very early days of nuclear power, tried to operate liquid-metal fast breeder reactors, but all experienced either accidents, other problems, or public opposition.
Thus, for a variety of reasons, the current mix of reactors around the world is heavily skewed toward light water designs. According to figures published in the March 2012 edition of Nuclear News, 351 of the 435 power reactors in the world are light-water reactors; 267 of them are PWRs and the other 84 are BWRs. Another 51 reactors are heavy water reactors. The remaining reactors include 17 carbon dioxide-cooled graphite reactors in the United Kingdom and 15 light water-cooled graphite reactors in Russia. Only one liquid metal-cooled power reactor is presently in operation, also in Russia. (These numbers include the reactors presently shut down in Japan for which no final decisions on future operation have been made.)
Today, the proponents of advanced designs argue that it is time for another look at some of the “new-old” technologies. Certainly, although the generic descriptions based only on coolant type sound similar in some cases, the details of many of the new concepts are very different. The prospects may be different as well.
The Very High Temperature Reactor, for instance, is designed for much higher temperatures than past and current gas-cooled reactors achieve, which would allow them to be used for high-temperature process heat applications—including hydrogen production—as well as for electricity production. The Gas-cooled Fast Reactor is designed to use a fast neutron spectrum, which was not the case for the early reactors. Liquid-metal concepts being developed include concepts generally similar to older designs, as well as concepts such as the “traveling wave reactor,” that are radically different. The Supercritical Water-Cooled Reactor would operate at the critical point of water, giving it certain advantages over either pressurized or boiling water reactors. There are also other “advanced” light-water concepts, such as an integral small reactor design, that are under development in several laboratories and companies.
While early versions of many of these designs may not have succeeded in the marketplace, advocates for these technologies believe the situation is different now. Technological advances and new requirements may result in a different mix of designs in the future. What history teaches us, they would argue, is that it is not possible to predict the technology “winners” at the outset of the effort.
Therefore, it is interesting to speculate on how this new initiative to develop more advanced designs may play out. There are already strong pressures to focus on the integral light-water design; based on well-understood light-water technology, the argument goes, such designs will be much easier to develop and license. That is true, and is a strong argument for developing integral small reactors to meet near-term needs.
In the longer term, however, some of the non-light water reactors could ultimately achieve greater levels of passive safety, efficient fuel utilization, economic performance, and proliferation resistance. Perhaps the biggest message history has to offer is that we should not declare victory and cancel R&D on other reactor design options as soon as the next new reactor design begins operation. We need to continue to explore multiple options to produce designs that can continue to meet evolving needs. In spite of the enthusiasm of the proponents of each design, we can expect that not every design option will achieve the full potential anticipated at this point. I am fond of citing Admiral Rickover, who famously observed that real reactors never achieve the performance predicted for them on paper.
Invoking Rickover brings up a key difference between the era of nuclear firsts and today. The early history of nuclear power was dominated by weapons development and the performance demands of naval propulsion. The global competition that was under way dictated a lot of the choices that were made in that era. The forces driving the development of the Gen IV reactors don’t have the same power to secure enormous budgets and generate a national consensus.
Of course, these nuclear development initiatives will take place in parallel with further development of other energy technologies, including renewables, clean coal, and carbon sequestration, and the ultimate mix of technologies will be determined by the marketplace. My guess is that it will be a long time before the advantages of nuclear power for baseload capability will be matched by renewables. Gen IV reactors, then, are likely to be crucial part of the global energy mix in this century. It's vitally important that we get this right, and that means looking back toward the first era of nuclear power as well as forward into the future.