Advanced Reactors: Proliferation and Terrorism Concerns
Edwin S. Lyman
Senior Scientist, Global Security Program
Union of Concerned Scientists, 1825 K St, NW, Ste. 800
Washington, DC 20006
DRAFT: September 19, 2018
Introduction
The workhorse of the commercial nuclear power fleet in the United States and most other nations is the light-water reactor. Over the last few years there has been renewed interest in reactors that use coolants other than water. Such reactors—often referred to as “advanced” despite the ancient vintage of many of the concepts—include liquid-metal-cooled, gas-cooled, and molten-salt-cooled designs. Articles frequently appear in the popular press lauding the next generation of nuclear power plants and the bold startup companies that hope to bring them to market. Think tanks, lawmakers, and the Energy Department have put forward optimistic visions of the role that non-light water-cooled reactors could play in mitigating climate change and addressing the nuclear waste disposal problem. Some private companies hope to deploy their first units by the late 2020s, and the Energy Department hopes to build its own liquid metal-cooled fast spectrum reactor by 2025.
However, there has been little discussion of the impacts of commercial deployment of these reactors on the risks of nuclear proliferation and terrorism. Many designs would use high-assay low enriched uranium (HALEU). Some would require nuclear weapon-usable materials such as uranium-233 (U-233), plutonium and other actinides. And most advanced reactor fuel cycles involve reprocessing—that is, chemical processing of spent nuclear fuel to extract re-usable nuclear materials. While some fuel cycles are of greater concern than others, all would introduce significant challenges for safeguards and security. For instance, although HALEU is not considered a direct-use nuclear weapon material, it has a higher security classification than lower-enrichment grades of LEU. This has not been a major problem to date because worldwide demand for HALEU has been limited. Detailed analysis is needed of the security implications of building a global infrastructure for production, transport, storage, and processing of large quantities of HALEU. Also, the difficulty of protecting and accurately accounting for the large flows of weapon-usable materials that would be separated by reprocessing plants in these schemes, both at and away from reactor sites, should be assessed. Some issues of concern are discussed below.
High-assay Low Enriched Uranium (HALEU): Security and Safeguards
Many of the advanced reactor designs being discussed today would use HALEU, with enrichments greater than 5% and less than 20% uranium-235 (U-235), instead of the highly enriched uranium (HEU) or mixtures of plutonium and uranium that such designs would have used in the past. HALEU would be necessary to provide sufficient fuel reactivity and allow for extended fuel burnups. Proposed reactors that would use HALEU of varying enrichments include the Oklo fast reactor (2 MWe); the ThorCon thermal molten salt reactor (250 MWe per module); and the X-Energy pebble-bed high temperature gas-cooled reactor (75 MWe per module).
The amount of HALEU that a large power reactor would require is huge compared to current supply and demand. For example, a 4-module, 1000 MWe ThorCon plant would require about 9.5 metric tons (MT) of 19.5%-enriched HALEU for the initial cores, and a supply of 3.8 MT per year.[1] In contrast, current worldwide demand for U.S-origin HALEU is around 1.5 MT per year, and the excess HEU stockpile that the DOE has designated to be blended down to meet that demand will be exhausted by 2040.[2]. Additional ways for the DOE to obtain HEU from existing stockpiles for blending down to HALEU, such as a plan to build a facility at the Idaho National Laboratory to recover HEU from irradiated naval reactor fuel, are speculative at this point. In 2015, the DOE assessed this option would be “very high cost” and “very high risk.”[3] Assuming there is only about 30 MT of 19.5%-enriched HALEU equivalent (1.5 MT x 20) remaining, a 1000 MWe ThorCon plant would consume the entire amount in less than six years.
Other proposed advanced reactors would also require tons of HALEU per year. X-Energy, said to have a thermal efficiency of 38%, would use 15.5% enriched uranium at a burnup of 163,000 megawatt-days per metric ton heavy metal (MWd/MTHM), and that a 7-gram pebble would generate 27.4 MW-hours of thermal energy. Thus a 1000 MWe (2632 MWth) X-Energy plant would require 841,500 pebbles containing 5.9 MT of HALEU per year. And although there is little public information about the Oklo 2 MWe design, a similar reactor concept, the Los Alamos “Megapower” reactor, would use 2.6 metric tons of HALEU for a 5-year core lifetime, which works out to 260 metric tons per GWe-year.
Thus the HALEU demand for a reasonably sized fleet of advanced reactors such as ThorCon or X-Energy could easily be two orders of magnitude greater than the current supply rate, exceeding what could be obtained from additional HEU downblending and requiring new production. An industrial-scale HALEU fuel cycle to support conversion, enrichment, fabrication, transportation, waste management, and (hopefully) disposal would need to be built from scratch. And if the U.S. moves forward with commercialization of HALEU-fueled advanced reactors, U.S. companies may seek to export them, and the rest of the world may pursue their own programs. In light of this, it is essential to evaluate the risks of a greatly expanded civil use of this material, both in the U.S. and worldwide.
What are the proliferation and terrorism risks of commerce in HALEU? With regard to the former, there is no formal distinction between high-and low-assay LEU in IAEA safeguards. The IAEA classifies all LEU, including HALEU, as “indirect use material,” and the agency’s goals for timely detection of diversion are the same as for lower-assay LEU.
Nevertheless, the additional proliferation risk of stockpiling HALEU has been recognized by the international community. Under the November 2013 Joint Plan of Action, Iran committed to voluntarily reduce its inventory of 20% (actually, “up to 20%”) enriched LEU to only what was needed for “working stock” for its research reactor, and to temporarily not enrich above 5%; measures that were strengthened in the July 2015 Joint Comprehensive Plan of Action (JCPOA). These agreements have created a new de facto safeguards category for HALEU that acknowledges its potential for reducing the time to acquire a nuclear weapon, at least for a country that possesses uranium enrichment facilities.
And the security sector appreciated the risks of HALEU long before that. The categorization table in INFCIRC/225, the IAEA’s recommendations for protecting nuclear materials from terrorist theft, classifies 10 kilograms of U-235 contained in LEU enriched to 10% U-235 or above (sometimes referred to as “intermediate-enriched”) as Category II, the second most sensitive security category. In contrast, any quantity of LEU under 10% U-235 enrichment is classified as less sensitive Category III material. This 10% enrichment threshold was present in all revisions of INFCIRC/225 going back to Rev.1 in 1975.
In 1979, the U.S. Nuclear Regulatory Commission (NRC) amended its regulations to impose physical protection requirements on Category II and III material (which it called “special nuclear material of moderate and low strategic significance,” respectively). In doing so, it largely adopted the INFCIRC/225/Rev.1 categorization table, and in particular preserved the Category II threshold of 10 kilograms for intermediate-enriched LEU. The NRC’s rationale for the amendments was both to enhance “domestic protection of such materials…” and to demonstrate “U.S. willingness to accept international security standards.” [4] The NRC pointed out that [5]
“In regard to … LEU … clandestine enrichment to higher levels may go beyond the capability of subnational terrorists, but it does not go beyond the capability of other governments. Unless properly safeguarded, low enriched uranium could be stolen on behalf of foreign governments and enriched to explosive useable levels after it is smuggled out of the U.S.”
Thus the threat the rule sought to address was a subnational group with state support.
However, the Statement of Considerations for the rule did not explain why it treated HALEU differently depending on whether it contained more or less than 10% U-235. The obvious reason is that fewer separative work units (SWU) would be needed to produce HEU from the more enriched material. According to a U.S. government official, the thinking was that it was “much easier” to use the material with U-235 > 10%, but that was at a time when enrichment was “a bit slower” than today. More elaboration on this point was provided in a 1998 paper by Forsberg:[6]
“Uranium-235 with enrichments between 10 and 20 wt % are not weapons-usable, but could be converted to weapons-usable materials with a relatively small uranium-enrichment plant. The complexity of these enrichment plants is such that this could not be accomplished by a subnational group, but it could be accomplished by many countries. The third category is uranium enriched to .ess than 10 wt % U-235 but above natural enrichment (0.71 wt % U-235). To convert this material to weapons-usable material, a substantial uranium-enrichment plant would be required. Such a plant would involve massive resources and would be very difficult to hide.”
But how important is this distinction in today’s threat environment? The difference in the quantity of SWU needed to produce enough HEU for a weapon doesn’t appear that significant. The goal of the NRC’s security regulations is to prevent theft of a “formula” quantity of strategic special nuclear material (e.g. HEU or plutonium), or to deter theft of smaller amounts that could be combined to form one formula quantity. Production of one formula quantity of 90% enriched HEU (about 5.5 kilograms) starting from 5% enriched LEU feed would require about three times more SWU than using 19.75% enriched feed, and 1.7 times more than using 10% enriched feed. This may be a significant difference for a gaseous diffusion plant, but it is not necessarily one for a modern gas centrifuge plant.
Nevertheless, the NRC staff maintains that the 10% threshold is still relevant. In the January 2015 final regulatory basis for a proposed rule on “enhanced security for special nuclear material,” the staff proposed strengthening security on LEU enriched to 10% or above because it believed that “greater oversight is required,” based on risk insights developed since the 9/11 terrorist attacks. One may read into this statement the NRC staff’s judgment that the capabilities of subnational groups to enrich uranium have increased—either with state assistance or independently.
But even that assumption doesn’t fully explain the 10% threshold. If enrichment by subnational groups is more plausible today, then lower assay grades of LEU would also need greater protection.
One way to look at this is to posit that if terrorists actually were able to develop a “quick and dirty” enrichment capability, they would seek to build the smallest possible centrifuge cascade and thus would want to obtain the most highly enriched feedstock available. (The underlying assumption is that if adversaries had access to a large enough enrichment capacity, their preference would be to use natural uranium feed.) According to Glaser, the smallest practical first-generation cascade would have 60 centrifuges at 2 SWU per year each and a total capacity of 10 SWU per month, for a total of 120 SWU per year.[7]
At 120 SWU/year, adversaries’ options for enrichment would be limited. But they could produce one formula quantity of weapons-grade HEU (> 90% U-235) in one year starting with uranium enriched at 15%, the midpoint of the 10-20% range, assuming a 0.5% tails assay (characteristic of early nuclear weapons programs). In contrast, 270 SWU would be required to achieve the same goal with 5% enriched feed. Similarly,120 SWU could be used to produce one formula quantity of 40% enriched HEU from 10% enriched feed, whereas 220 SWU would be needed if 5% enriched feed were used. Therefore, a requirement for additional physical protection for intermediate-enrichment LEU is consistent with its potential for enrichment to a formula quantity of HEU within one year using a minimally capable centrifuge plant.
In addition to requiring additional physical protection measures for intermediate-enriched HALEU, the NRC also imposes more rigorous material control and accounting measures. However, as mentioned above, this is not the case for IAEA safeguards. The fact that the IAEA regards all LEU as equivalent with respect to safeguards, yet recommends different security measures for LEU with enrichments below and above 10%, seems to be an inconsistency.
If the production and civil use of HALEU is to be greatly expanded, these inconsistencies need to be resolved. If intermediate-enriched HALEU requires more stringent measures than lower-assay LEU, then it warrants more rigorous safeguards as well. U.S. government agencies and the IAEA should take a hard look at the safeguards implications of a commercial HALEU fuel cycle in view of contemporary adversary capabilities, and adjust their requirements and recommendations accordingly. The IAEA should consider a smaller significant quantity (perhaps 50 kg instead of 75 kg) and a shorter timeliness detection goal (perhaps 6 months instead of one year) for HALEU with U-235 10% or greater. But such radical changes are nearly impossible at the IAEA. Short of that, the IAEA could consider separate tracking of HALEU production and use, similar to its approach for the “alternate nuclear material” neptunium-237, at least until agreement could be reached on more stringent measures. The U.S. could also impose stricter requirements on HALEU material accountancy and control in countries subject to 123 agreements for nuclear cooperation.
Another criterion relevant to proliferation is the quantity of SWU required to generate a unit of electricity at reactors that use HALEU. In general, the fewer SWU needed to support civil nuclear power, the better for nonproliferation. Transitioning to a fuel cycle that requires less enrichment could reduce the need for construction of new enrichment capacity around the world. However, many advanced reactor designs are not optimized for more efficient utilization of enrichment services. For example, a 1000 MWe light-water reactor with a 60-year lifetime would require about 150,000 SWU/GWe-year on average. In comparison, consider the ThorCon plant, which would use an average of 5 MT of 19.5% enriched HALEU per year (taking into account the initial core load and the eight-year lifetime of the reactor). Production of this fuel would use 200,000 SWU/GWe-year on average, or 33% more SWU/GWe-yr than an LWR. And an Oklo-type fast reactor would require a whopping 10 million SWU/GWe-yr. This is going in the wrong direction.
There is one potential application of HALEU in advanced reactors that could be beneficial for nonproliferation: its use as a driver fuel in once-through “breed and burn” reactors, such as the TerraPower “traveling-wave” reactor. In theory, such reactors could use uranium more efficiently in a once-through cycle by increasing the internal conversion of U-238 to Pu-239, and subsequent fission of Pu-239, without the need to reprocess and recycle spent fuel. However, such reactors would need fuels able to achieve very high burnups. Higher-enriched fuels such as HALEU would be needed for high-burnup driver fuel at startup (and until the system attained an equilibrium state where it could operate only on depleted uranium feed). If such systems were possible, they could reduce the quantity of SWU needed per GW-year. For example, TerraPower estimates the lifetime-averaged separative work for its reactor would be 30,000 SWU per GWe-yr, about 20% of that required for an LWR.[8] Unfortunately, major technical obstacles remain, including the development of fuel materials that can safely withstand very high burnups.
HALEU: The Radiological Weapon Threat
Apart from its attractiveness for a nuclear weapons program, HALEU also presents a radiological threat because it could be used to construct a stealthy and potentially devastating radiological emission device (RED). This danger was highlighted by a September 1999 criticality accident at the JCO uranium conversion plant in Tokai-mura, Japan. The accident was caused when workers, circumventing nuclear criticality controls, poured 16.6 kilograms of 18.8% enriched uranyl nitrate solution (containing just over 3 kilograms of U-235) into a precipitation tank which had a batch limit of only 2.4 kilograms of total uranium of that enrichment. The tank, with a diameter of 45 centimeters and a height of 60 centimeters, was surrounded by a cooling jacket through which water was circulated. The error turned the precipitation tank into a critical aqueous solution nuclear reactor, the cooling jacket serving as both a neutron reflector and a heat removal mechanism. After a number of high-power pulses, the system attained a steady state at a power of around 1 kilowatt-thermal and ran for nearly twenty hours before personnel were able to fully drain the cooling jacket and render the reactor subcritical.
The accident exposed three workers in the immediate vicinity of the tank to doses of 3, 10, and 17 Sieverts (Sv). The two most highly exposed workers received lethal doses (e.g. exceeding about 4 Sv) and eventually succumbed to acute radiation syndrome. But the unshielded reactor also generated steady-state neutron and gamma radiation fields that combined were more than ten times background (3 µSv/hr) at 750 meters from the reactor. At 50 meters from the reactor, the neutron dose rate was 10 mSv/hr, and the gamma dose rate about ten times lower. But the brief initial pulse was one thousand times more intense, resulting in a dose rate that must have been on the order of 10 Sv/hr at 50 meters and 3 mSv/hr at 750 meters.
Although this was an unfortunate accident, the concern is that it could be deliberately reproduced by an adversary in possession of a quantity of HALEU on the order of kilograms. The IAEA rightly observed that “from a nuclear engineering point of view, the water-cooled tank system was a very simple system.”[9] In fact, the solution reactor concept is so simple that it was the basis for the third and fourth nuclear reactors ever built, the 1943 and 1944 “water boilers” at Los Alamos, which used 14%-enriched uranium. With beryllium and graphite reflectors, the critical mass was as low as 565 grams of U-235 (4 kg of total uranium).
And while the off-site radiological consequences of the JCO accident were limited, the reactor itself was far from an optimized system. A higher power steady state could have been achieved with less nuclear material and a greater heat removal rate than the natural circulation in the cooling jacket provided. Historical “water boilers” were designed to operate at powers as high as 50 kW-thermal with only 5 kg of uranium enriched to 12.5%. Such an assembly, placed in a refrigerated truck with gamma shielding, and parked on a city street, might not be detected for days or longer, by which time tens of thousands of people could have been exposed to extremely high radiation levels.
The threat posed by such a scenario should be considered in developing appropriate physical protection measures for the commercial use of HALEU. Given the relatively small quantities of HALEU that could be used to fuel a critical system, it may be warranted to reduce the Category II lower limit of 10 kg U-235 for HALEU with 10% or greater U-235.
Advanced Reactors, Reprocessing, and Weapon-usable Fuels
Many advanced reactor designers assert that their systems would be capable of “consuming” spent fuel from light-water reactors and recycling their own nuclear waste. These features have become a major selling point for the public and for many on Capitol Hill. But vendors and their supporters rarely explain what such operations would entail: reprocessing and fabrication of plutonium and other transuranic elements into fresh fuel. They avoid discussing the safety and environmental risks of these processes, not to mention the security risks associated with large-scale plutonium separation and widespread deployment of plutonium-fueled reactors. And even though the approaches needed to adequately safeguard and protect these reactors and their fuel cycles have not even been defined for the most part, there is a major push to deploy advanced reactors within the next decade both in the United States and abroad.
For example, a number of liquid metal-cooled fast reactor developers are pursuing construction of test or demonstration units. In the U.S., Oklo began pre-licensing discussions with the NRC in 2016. Also, the DOE has $35 million in FY 2018 funding to begin studies of the “Versatile Test Reactor,” a 300 MWe sodium-cooled fast reactor. And several U.S. vendors are seeking to build plants in Canada, which is becoming a sort of Cayman Islands for off-shoring advanced nuclear plants because of a widespread (but probably unjustified) belief they will be easier to license there. For example, the company Advanced Reactor Concepts recently announced a collaboration with New Brunswick Power to explore “the potential future deployment of the ARC-100,” a 100 MWe sodium-cooled fast reactor, “at NB Power’s Point Lepreau nuclear plant site and thereafter at other sites in Canada and worldwide.”[10]
Although little is publicly known about these three projects, they are all metal-fueled fast reactors modeled after the defunct Experimental Breeder Reactor-II (EBR-II). And all apparently intend to use plutonium and other transuranic-based fuels at some point. Although Oklo initially plans to use HALEU, CEO Jacob DeWitte testified before Congress in July 2017 that “…our reactor can actually consume the used fuel from today’s reactors as well as the depleted uranium stockpiles around the nation. In fact, fast reactors like ours could power the world for over 500 years with the global inventory of used fuel and depleted uranium. Our reactors can also assist with plutonium disposition by consuming excess cold war era materials and turning them into clean, peaceful energy.”
Also, the DOE has chosen a metallic alloy of plutonium and 5% enriched LEU fuel as the preferred baseline for the Versatile Test Reactor.[11] And the ARC-100 would reportedly use HALEU, for its initial core, but after a 20-year cycle would send the spent fuel for pyroprocessing (a type of non-aqueous reprocessing) to extract plutonium and other transuranics for fabrication into fresh fuel.
Why is this a problem? Aside from the proliferation and terrorism risks associated with reprocessing and plutonium fuel fabrication facilities, DeWitte also testified that Oklo hopes to sell its “microreactor” to “remote and rural communities.” Similarly sized heat pipe-cooled fast reactors require about 2.5 MT of fuel enriched to about twenty percent; thus each reactor could require several hundred kilograms of plutonium.[12] Who is going to provide the necessary security, and at what cost? And how will the material be safeguarded against diversion? In non-nuclear weapon states, verification of plutonium-fueled reactors widely dispersed throughout remote communities, such as the Arctic, could prove an enormous burden for the IAEA.
One may wonder why so many advanced reactor startups emphasize reprocessing and spent fuel recycling. It may be a legacy of message testing that dates back to the Global Nuclear Energy Partnership (GNEP) era, when the DOE had proposed a global system for supplying fresh fuel and reprocessing spent fuel. In 2006, Bisconti Research was commissioned by Idaho National Laboratory to do focus group testing for GNEP communications.[13] The study found that the top message to emerge from the focus group was “With GNEP, the U.S. joins other nations recycling fuel and will recycle valuable fuel to produce energy and simplify waste management.”
But reprocessing and spent fuel recycling are not essential for many types of advanced reactors, including liquid metal-cooled fast reactors, which can operate on HALEU fuel on a once-through cycle. For molten salt reactors, however, on-line processing of the liquid fuel is inescapable. At a minimum, gaseous and noble metal fission products have to be continuously removed. And in order to achieve high burnups, neutron-absorbing fission products would also need to be periodically extracted from the fuel. The difficulties in developing material accounting methods for the required on-site reprocessing plants are obvious and widely appreciated in the technical community.
To illustrate those challenges, consider the Transatomic Power (TAP, which ceased operations very recently) molten fluoride salt reactor. The TAP reactor is a 520 MWe spectral-shift reactor, fueled by LEU. During the first part of its 29-year operating cycle, an intermediate-energy neutron spectrum promotes conversion of U-238 to plutonium. In the second part of the cycle, a thermal spectrum promotes fission of the plutonium that has built up in the core. Consequently, the plutonium content of the molten salt rises steadily to a peak of about 4 MT after 20 years of operation, after which it slowly decreases to about 3 MT at shutdown.[14] In order to achieve a 29-year lifetime, the fuel must be processed to remove neutron-absorbing rare earth fission products on an approximately 50-day cycle, or 7.3 cycles per year. When the core plutonium content is at its peak, the plutonium throughput of the processing system would be about 29 metric tons per year for a 520 MWe reactor. This is an enormous flow of plutonium given such a small reactor. In comparison, the peak throughput of a large commercial reprocessing plant is around 8 MT of plutonium per year, which corresponds to the annual spent fuel discharges from about forty 1000 MWe LWRs.
Because there is so little known about this reactor and the required fuel processing system, not much can be said specifically about it. It is not even known if the process is feasible or even possible. However, it is difficult to see how an effective safeguards approach based on material accountancy could be developed for this reactor. There is little information available about important factors for material accountancy, such as uncertainties in core inventory calculations, separation efficiencies, waste streams, and process holdup (material stuck in the system). In 2014, a review article pointed out that “acquisition of fundamental data for the extraction processes is still needed especially for the actinide-rare earth fission products separation.”[15] Subsequently, experimental work on separating uranium and neodymium from molten fluoride salt found only “low” extraction efficiencies, calling into question the current processing approach.[16] Even if a process loss rate of 0.1% per year could be achieved, which would be remarkable, this would correspond to more than one significant quantity of plutonium (8 kilograms) going into waste every year. How will this material be measured?
One approach for safeguarding such facilities would be to greatly reduce reliance on material accountancy and to rely instead on containment and surveillance (C/S) and process monitoring, as has been proposed for fast reactor fuel pyroprocessing plants. But these alternative measures have inherent limitations and are unable to provide the same level of confidence as material accountancy. A shift in philosophy by the IAEA to accept such measures as substitutes for material accountancy would be a step in the wrong direction, ultimately increasing the risk of proliferation. The IAEA should hold its ground and not accept weaker safeguards standards for facilities that may well be unsafeguardable in practice. Also, in addition to its impact on international safeguards, the inability to conduct timely material accountancy would hamper the ability of state authorities to rapidly assess the validity of theft allegations. If someone were to call in a threat, claiming to have diverted enough plutonium for a bomb from a processing facility, it could take weeks or longer to establish whether material was actually missing.
In summary, the threats of nuclear proliferation and nuclear terrorism must be primary considerations in charting the future course of nuclear energy. New nuclear technologies should only be deployed only if they can be effectively safeguarded from nuclear material diversion and protected against theft. To this end, the U.S. should make all advanced reactor and fuel cycle demonstration facilities, such as the Versatile Test Reactor, eligible for IAEA safeguards, and engage the IAEA on developing safeguards during the facility design phase. Congress should also ensure that all demonstration reactors on DOE sites be subject to NRC licensing, so that the NRC can develop methods for oversight of material accounting and physical protection at advanced reactors and fuel cycle facilities.
Development of once-through breed-and-burn reactors that could increase uranium utilization and reduce the need for enrichment would undercut the major rationales for reprocessing and plutonium recycling, helping to further nonproliferation goals. However, such reactors present many technical challenges. At present, the LWR operating on a once-through cycle still appears to be the least risky option for future nuclear power deployment.
References
- J. Devanney, ThorCon, personal communication with the author. May 4, 2018.
- U.S. Department of Energy, Tritium and Enriched Uranium Management Plan Through 2060, Report to Congress, October 2015.
- Ibid.
- U.S. Nuclear Regulatory Commission, “Special Nuclear Material of Moderate and Low Strategic Significance: Safeguards Requirements,” proposed rule, 43 FR 22216, May 24, 1978.
- U.S. Nuclear Regulatory Commission, “Safeguards Requirements for Special Nuclear Material of Moderate and Low Strategic Significance,” final rule, 44 FR 43280, July 24, 1979.
- C.W. Forsberg et al., “Definition of Weapon-Usable Uranium-233,” ORNL/TM-13517, Oak Ridge National Laboratory, March 1998.
- A. Glaser, “On the Proliferation Potential of Uranium Fuel for Research Reactors at Various Enrichment Levels,” Science and Global Security 14 (2006) 1-24.
- J. Gilleland, J., R. Petroski and K. Weaver, The Traveling-Wave Reactor: Design and Development,” Engineering 2 (2016) 88–96.
- International Atomic Energy Agency, Lessons Learned from the JCO Criticality Accident in Japan in 1999, undated.
- “ARC Nuclear Announces Agreement to Work with New Brunswick Power and Its Affiliates to Explore Commercialization of the ARC Advanced SMR in New Brunswick, Canada,” press release, Business Wire, July 9, 2018.
- E. Lyman, “The Versatile Test Reactor Debate: Round 2,” All Things Nuclear blog post, June 11, 2018. www.allthingsnuclear.org.
- P. Mcclure et al., “Design of Megawatt Power Level Heat Pipe Reactors,” LA-UR-15-28840, Los Alamos National Laboratory, November 2015.
- Bisconti Research, Inc. “Focus Groups to Test GNEP Communications,” Boise, August 2006.
- B. Betzler et al., Assessment of the Neutronic and Fuel Cycle Performance of the Transatomic Power Molten Salt Reactor Design,” ORNL/TM-2017/475, Oak Ridge National Laboratory, September 2017.
- J. Serp et al., “The Molten Salt Reactor (MSR) in Generation IV; Overview and Perspectives,” Prog. Nuc. Energy 77 (2014) 308-319.
- D. Rodrigues, G. Durán-Klie, and S. Delpech, “Pyrochemical Reprocessing of Molten Salt Fast Reactor Fuel: Focus on the Reductive Extraction Step,” Nukleonica 60 (2015) 907-914.