From “Inherently Safe” to “Proliferation Resistant”: New Perspectives on Reactor Designs

Abstract

A recent American “mini-series” on Chernobyl, widely watched across the world, presented viewers with the concluding finding that this massive accident had occurred because the reactor design had inherent flaws; flaws that were known but not previously fixed because it was “cheaper” that way. The reactor design in question is the RBMK, and this paper will argue that this design was far from “cheap,” neither then nor now, and that its adoption as the second standard design for the Soviet Union’s nuclear power reactor fleet was based on much more than economic considerations. With the benefit of hindsight, it is easy to forget that reactor designs are always chosen for a multitude of reasons and never solely based on their technical or economic merits. Based on archival research, interviews, and industry publications, I show that approving and building RBMK reactors made good sense at the time, despite later claims to the contrary. Then I take the examples of a small modular reactor (SMR), the proposed NuScale Power Module, and a fast neutron reactor, TerraPower’s proposed Traveling Wave Reactor, to argue that we witness comparable negotiations today, as new designs for reactors (1) attempt to fit into existing safety and regulatory frameworks, (2) navigate security and nonproliferation concerns, and (3) embody visions of a specific sociotechnical order. I conclude that technical designs never occur in a socioeconomic, political, or cultural vacuum; instead, they are developed by people steeped in social norms, regulatory concerns, and economic expectations of a specific time and place. In the spirit of making this point relevant to practitioners, I will suggest ways of making these implicit frameworks visible, to actively and consciously start tweaking them, while staying aware of the implications that technical choices may have on our social expectations and vice versa.

Keywords:

• Reactor design
• regulation
• nonproliferation
• sociotechnical imaginaries

We are approaching a tremendous revolution in human life with which nothing hitherto experienced can be compared. It will not be long before man will have atomic energy at his disposal as a source of energy that will make it possible for him to build his life as he pleases.

—Vladimir Vernadskii, Head of the State Radium Institute, Leningrad, 1922

I. INTRODUCTION

Designing nuclear reactors is no straightforward business. Instead, design decisions are, and have been, the product of complex and often competing constraints and demands, including economic parameters, questions related to energy and national security, and available technical options. Yet designers rarely reflect on this environment that they are part of and that is shaping their choices. Also, as professionals trained in highly specialized disciplines, they typically don’t question the role of nuclear technology in society, other than perhaps lamenting a demanding regulatory regime and the lack of state support for novel design ideas. This paper argues that reactor design choices are never solely based on their technical or economic merits. Rather, each design is shaped by, and simultaneously shapes, its socioeconomic, political, and cultural context.

Central notions, such as safety, security, and nonproliferation, have accompanied the development of nuclear reactor technologies from early on. This paper adds the concept of “sociotechnical imaginaries,” developed by scholars in the field of Science and Technology Studies (STS), to illustrate how different technological choices embody, and embed, visions of a specific sociotechnical order.^1,2 By making the connection between technical choices and sociotechnical visions explicit, this paper seeks to provide tools to practitioners to see these connections more clearly, including their own role as engineers, planners, or policymakers, and to situate their choices in a more comprehensive way. For design choices, consciously or not, express regulatory preferences, suggest governance options, and shape the decision-making process.

Back in the 1950s, once the idea of controlled nuclear fission gained traction with engineers, there seemed to be no limit to their imagination. Early designers experimented with a variety of materials to cool and moderate their machines, with shapes of fuel elements, sizes of cores, numbers of loops, and many other parameters. Engineers built models, prototypes, and sometimes research reactors. In most countries, however, this variety soon thinned out and only a few, arguably the most promising, designs received further support and funding. The Geneva conferences on the “peaceful uses of atomic energy” helped this process along by showcasing different options and reaching international consensus on a preferred design for large, commercial power reactors, at the time the pressurized light water design that became known as PWR in the West and VVER in the USSR.^a

Other designs, including the boiling water reactor (BWR) design, continued at a more moderate pace, and in specific niches, such as Soviet submarine propulsion, more exotic design varieties persisted. While the United States developed PWRs and BWRs as the main types for large commercial nuclear power plants, and soon distributed these designs to other nations through the Atoms for Peace program, this was not the only path nuclear reactor designers chose. In Britain, engineers favored a gas-cooled design; in France, a design that produced weapon-grade plutonium was abandoned as a commercial option only years into an expensive development program.³

In the Soviet Union, reactor designers continued experimenting with different designs into the late 1960s. In 1956, Igor V. Kurchatov, prominent physicist and scientific leader of the Soviet atomic bomb project, mentioned “up to ten” designs that Soviet scientists and engineers were working on.^4,5 Scientists argued that only by developing in many directions and experimenting with a variety of design options, coolants, and cooling schemes could they produce a rich and diverse scientific and technical knowledge base for the nuclear industry.⁶

Viktor A. Sidorenko, one of the chief developers of the VVER, presented the different reactor design lines in the form of branches of a “tree” in a 1997 paper.⁷ In this timeline he tried to trace the major directions in designing power reactors over the course of Soviet history. What is most remarkable from today’s perspective is the enduring diversity of designs. Eventually, however, as in other nations, the Soviets also picked a couple of designs for commercial development. Manufacturing bottlenecks on the one hand, and experience with military reactors on the other, enabled specific pathways that led to the adoption of both the VVER and the RBMK (more on this reactor type below).⁶

Sidorenko’s attempt at reconstructing the various Soviet design lines relates nicely to what Bijker et al. called the “social construction of technology” in the 1980s (Ref. 8). They observed that multidirectional models that map the actual diversity of design directions are often streamlined in retrospect, leading to the (false) impression of a linear development of one technology that succeeded because it was “the best” (or the cheapest or the most functional). They argued that the outcome—what model or design eventually makes it—depends on a number of factors, including what problem it promises to solve and whether enough “relevant social groups” consider the solution it offers as stable “closure” to the problem they perceive to exist in the first place. Their example was the bicycle, but the approach can accommodate reactors just as well. In this paper, I consider three pairs of problems and solutions offered by specific reactor designs. First, the problem of safety, and the promise of inherent safety; second, the problem of proliferation (that is, the use of materials or know-how for a weapons program) and the idea of proliferation resistance; and third, the problem of purpose, the guiding vision, or imaginary for the way a new design fits with society, that a specific design promises. The first two problems have come to dominate our thinking about nuclear technologies, but they weren’t always perceived as fundamental concerns affecting nuclear energy, at least not in the way we understand them today. The third problem is perhaps most vexing of all, as the vision embedded in a design typically remains tacit, even while it defines which solutions to the other two problems become possible. By making a design’s guiding vision discernible, the concept of sociotechnical imaginaries can provide a new tool for practitioners to situate, adapt, and justify design choices in a comprehensive and nuanced way.

Design decisions, once made, cannot be easily reversed. Building nuclear reactors requires significant resource commitments, from forging capacity to quality materials, from precision instrumentation to fuel manufacturing, from training skilled workers to educating technical specialists, and from crafting regulatory guidelines to implementing rules and verification processes. Large sociotechnical systems support each design, systems that take time to set up and that in turn require a continuous production of reactors to maintain themselves.^9,b Once in place, these systems are difficult to modify.³ Historian of technology Thomas Hughes maintained that as systems mature, they acquire momentum and style, and it becomes increasingly harder to intervene and change direction.^{c, 9} Beyond the immediate reactor design, such systems even constrain future choices. As my colleague Aditi Verma has recently pointed out, today’s reactor designers work in a predefined space where choices of materials and processes are at least in part determined by available, and sometimes constraining, regulatory frameworks.¹⁰ On the other hand, reactor design choices may not prevail in the long term, regardless of the strategy. One of the premier nuclear reactor construction companies in the United States, Babcock & Wilcox, quietly ended their small modular reactor (SMR) program in 2017 (11). The Russian nuclear energy strategy for many years included plans to maintain and even expand the RBMK line, and only in the mid-2000s shifted to replacing them with VVERs (12, 13, and 14).

It is important to remember that reactor design choices are never just technical decisions about what works best, nor are they simply economic decisions about what is the cheapest. Many demands play into these decisions, starting with energy policy more broadly, to anticipated or projected electricity demand, and considerations of military need. Then and now, designers and planners rarely assess nuclear power in isolation. Rather, they see it as one piece of the energy (and today, climate) puzzle, and thereby choose a vision for society along with a design for nuclear technologies. Sometimes, however, these visions and designs are disconnected. Announcements of nuclear energy providing “the cure” to every conceivable problem, from energy to transportation, agriculture, and medicine, while guaranteeing progress and modernization in general, smack of advertising and 1960s wishful thinking, rather than accurate predictions of a complex energy future.^d Other narratives have tried to address this complexity, for example, by invoking “inherent safety” and “proliferation resistance.” What the notion of sociotechnical imaginaries suggests is that designing nuclear reactors entails designing nuclear societies; by designing a reactor, its safety parameters, and its proliferation resistance, we simultaneously envision a world in which this reactor design works and works well.

This paper shows how this complexity can be broken down, using the concepts of safety and security/nonproliferation to examine three distinct reactor designs that represent both differences and similarities in technological choices, historical contexts, and cultural environments, and the sociotechnical imaginaries associated with each of them. My hope is that by mapping out a wide area, practitioners will be able to situate their own design(s) more comfortably, thoughtfully, and confidently on a spectrum of options rather than reproducing outdated and potentially misleading scripts from the past.

In Sec. II, I introduce my disciplinary roots in STS, situate this study in the relevant STS literature, and review the study’s research design. Section III explains the conceptual framework, followed  in Sec. IV by a discussion of three reactor designs and the explicit or implicit sociotechnical imaginaries associated with them. There, I show first that the design, construction, and operation of RBMK reactors made good sense at the time, despite later claims to the contrary. Secondly, I take the more current examples of the NuScale Power Module (an SMR) and of TerraPower’s Traveling Wave Reactor (TWR) (a fast neutron reactor) to argue that we witness comparable negotiations today, as new designs for reactors attempt to fit into existing safety and regulatory frameworks, navigate security and nonproliferation concerns, and embody visions of a specific sociotechnical order. In Sec. V, I return to the relevance of this type of STS analysis for nuclear practitioners.

II. RESEARCH DESIGN

This paper was first written for the conference, “The Nuclear and Social Science Nexus,” at the Nuclear Energy Agency (NEA) in December 2019. It is rooted in STS, a growing interdisciplinary field that draws on the full range of the humanities and social sciences to examine the mutual shaping of science and technology on the one hand, and our society, politics, and culture on the other. Scholars in STS have long engaged with nuclear energy; in fact, early controversies over nuclear safety were arguably part of the origin of the discipline when concerned scientists and engineers highlighted controversies among experts and advocated for more meaningful democratic engagement in consequential decision-making processes.^15,16 These early, sometimes activist STS studies on nuclear energy have provided a blueprint for studying other hazardous emerging technologies.^17–20

The STS scholarship on nuclear matters has since grown in scope and nuance, and direct collaboration between STS scholars and nuclear engineers is on the rise, both in research and curricular development.^{2,3,21–23,e} An STS perspective allows nuclear practitioners to include systematically questions of social context, both how this context shapes the questions asked and also how decisions in turn shape this context.

This paper draws on several traditions within STS, including social studies of technology, large technological systems, and sociotechnical imaginaries.^1,8,9,15 In my experience, these approaches are particularly relevant and accessible for engineers and practitioners. To illustrate how such STS concepts can be utilized in nuclear engineering, I chose three reactor designs: (1) the “Chernobyl-type” RBMK reactor designed by Soviet engineers in the late 1960s, (2) the NuScale Power Module, a small modular design nearing design certification in the United States, and (3) the fast neutron reactor proposed by TerraPower. These designs are very different in terms of technology, when they were developed, and in what kind of organizational structures, and I chose them deliberately to show that an STS lens is applicable and useful not just for one period or technology.

The problems I selected to highlight for each of these designs reflect central concerns with nuclear technologies today (safety, security, and nonproliferation), but by including a historical example I show that these concerns themselves developed over time and that their exact scope and meanings will continue to change. The STS concept of sociotechnical imaginaries allows me to expose the underlying sociotechnical vision of each design. I identify these visions by relying on archival research for the RBMK design and industry publications, regulatory documents, and media reports for the NuScale and TerraPower designs. For all designs, I also talked to designers and industry insiders or veterans.^f

Comparing these three designs means not only comparing three different technologies, but also three distinct sociotechnical imaginaries. The visions associated with the newer designs are still fluid, and the imaginaries I articulate here may not be the only possible ones or the ones that will succeed. And yet, by extending our scope of analysis from technical feasibility alone to the accompanying sociotechnical visions, we can begin to understand more clearly what implicit assumptions these designs embody, what they promise to accomplish beyond their technical merit, and whether these assumptions, or the designs, might need to be modified.

III. CONCEPTUAL FRAMEWORK

III.A. Safety

When developing the designs of currently operating reactors in the 1950s and 1960s, engineers focused on proving the technical feasibility of commercial reactors, but also on demonstrating their economic viability. Safety, even if the concept lacked today’s comprehensive scope, was part of this project; after all, power reactors utilized the technology first unleashed for powerful new weapons. Control over the fission process was literally at the core of this new form of energy. But the high initial cost for constructing a nuclear power plant proved at least as challenging as safely managing the nuclear chain reaction. For countries without nuclear weapons or significant natural resources, the high investment in nuclear energy eventually paid off, especially since they benefitted either from the American Atoms for Peace program that President Eisenhower initiated in 1953, or received technical assistance from the Soviet Union under their nuclear assistance program within the Council for Mutual Economic Assistance.²⁴ For countries like the United States and the USSR, however, such investments were quite another story. These countries were already spending enormous resources on their nuclear weapons programs and had plenty of alternatives available to produce heat and electricity for commercial purposes. Yet it was these two emerging superpowers that, ironically at the time of an intensifying arms race, became the main developers and exporters of commercial reactors. In those early years, then, engineers had to prove not only that nuclear reactors would work, but that these reactors could compete economically with more conventional power plants, such as hydro- or coal-powered stations.^6,25

The emphasis on safety, and eventually “inherent safety,” emerged only gradually as it became clear that accidents were possible, some with significant consequences.²⁶ In terms of design, this idea manifested itself in modifications of existing reactor types, not in entirely new designs. For example, engineers in the USSR kept upgrading the RBMK design, resulting in a series of RBMK “generations” that each featured safety improvements.⁷ Similarly, in the United States designers didn’t fundamentally alter their reactor lines, even though they continuously improved them based on experience with already operating units.^27,g The conversation about new reactor designs today revolves very much around the topic of safety, as sophisticated licensing, regulatory, and inspection regimes have evolved, albeit for only a few types of reactors.^h Designs that don’t fit into any of these regimes face additional challenges because the rules of what constitutes safety aren’t yet established for them. This has led designers to shy away from truly revolutionary innovation, relying instead on both a proven safety record of certain features and materials and innovating within the frameworks that were developed for existing designs.¹⁰

And yet, discussions about improved safety rarely reflect upon what the term might mean beyond strictly engineering parameters. Ignoring the myriad elements that make a large technological system safer, or less safe, reflects a technocratic attitude that assumes technical features alone can make a reactor safe. While there is a lot to be said for fixing technical problems, there is a risk in relying on technical modifications alone. Even though a lot of progress has been made in the nuclear industry to account for “human and organizational factors” to advance a system’s safety, there are still design proposals that use automation as a selling point, where human agency is seen as merely interfering with the technology, be it through regulatory constraints or operating errors.ⁱ Carefully weighing the pros and cons of human judgment in a variety of possible scenarios, including ones where technology malfunctions or fails, remains a contested domain.³⁰

III.B. Security and Nonproliferation

From the very outset, technical assistance programs such as Atoms for Peace were criticized for implicitly spreading the ingredients for a nuclear weapons program; materials, processes, and know-how intended for “peaceful uses” of nuclear energy could, critics argued, be diverted for more sinister purposes.^31–33 The Nuclear Nonproliferation Treaty of 1970 (officially the Treaty on the Non-Proliferation of Nuclear Weapons) tried to accomplish a balancing act between the promise of “peaceful” assistance and that of foregoing weapons programs. In 1974, India used a loophole in the treaty, that of “Peaceful Nuclear Explosions,” to test a nuclear device that used plutonium from civilian reactors supplied by Canada under a technical assistance program.³⁴ Over the years, and as more and more “cheaters” of the nuclear nonproliferation regime were caught, the risks of material security breaches and the proliferation of nuclear weapons increasingly caught the attention of reactor designers. The IAEA formalized processes to avoid proliferation by defining “proliferation resistance” as those characteristics of a nuclear system that impede “diversion or undeclared production of nuclear material, or misuse of technology.”³⁵ The agency further established barriers to proliferation that relate to materials, technologies, and institutions.³⁶

And yet, although nuclear engineers have increasingly become aware of proliferation concerns, the “proliferation resistance” of new reactor designs lags behind the issues that security studies scholars have identified as the most critical to address.³⁷ One reason lies in the stubborn segmentation of professional communities involved with nuclear questions. Engineers and designers are still predominantly trained to address safety and reliability, while international relations scholars, who more closely inform political decision makers, tend to rely on legal instruments to contend with hidden proliferation risks.^38,39 This disconnect sometimes leads to proposals that are not well thought out. For example, the threat of nonstate actors associated with “nuclear terrorism” has introduced questionable assumptions about the deterrence value of highly toxic materials (e.g., thorium producing spent nuclear fuel that’s allegedly too dangerous to handle safely).⁴⁰ Arrangements to increase proliferation resistance, such as fuel lease and take back, are being discussed without considering the legal complexities involved with actually implementing such plans. In most countries, spent nuclear fuel is defined as waste, and importing waste not produced domestically would require changing the law and amending constitutions.^41,42 And finally, proposals to site new power reactors below grade does more than allegedly increase their security. It also suggests that siting a reactor underground, “out of sight, out of mind,” would somehow placate the safety concerns of a skeptical public. Worse yet, such proposals can indicate a lack of thought devoted to the reactor’s end of life.^j After all, burying what remains of a nuclear reactor, its spent nuclear fuel, stands out as one of the most controversial and as yet unresolved issues of the nuclear industry today.⁴³

III.C. Sociotechnical Imaginaries

The concept of sociotechnical imaginaries was first proposed by Sheila Jasanoff and Sang-Hyun Kim in 2009 and refined in their 2015 edited volume, Dreamscapes of Modernity, as “collectively held, institutionally stabilized, and publicly performed visions of desirable futures, animated by shared understandings of forms of social life and social order attainable through, and supportive of, advances in science and technology.”^1,2 In other words, sociotechnical imaginaries can be understood as visions of how new technologies will transform societies and how future societies will look like once they embrace a new technology. These imaginaries can originate in science fiction, or in a state-sponsored program. Once they gain traction, they compete with other imaginaries that may envision a very different world, and it may take institutions to select, promote, and ultimately implement them. In Jasanoff’s words:

Sociotechnical imaginaries . . . are not limited to nation states . . . but can be articulated and propagated by other organized groups, such as corporations, social movements, and professional societies. Though collectively held, sociotechnical imaginaries can originate in the visions of single individuals or small collectives, gaining traction through blatant exercises of power or sustained acts of coalition building. . . . Multiple imaginaries can coexist within a society in tension or in a productive dialectical relationship. It often falls to legislatures, courts, the media, or other institutions of power to elevate some imagined futures above others, according them a dominant position for policy purposes. Imaginaries, moreover, encode not only visions of what is attainable through science and technology but also of how life ought, or ought not, to be lived; in this respect they express a society’s shared understandings of good and evil.²

While we can identify sociotechnical imaginaries for most technological trajectories of the past and present, nuclear energy has carried forward a very explicit sociotechnical imaginary from the outset. Initially, this vision entailed a dramatic departure from nuclear weapons, a redemption of sorts for having unleashed the horrors of the atom bomb. Soon, this imaginary of a peaceful, nuclear-powered society was joined by efficiency (electricity “too cheap to meter”), and boundless, continuously availably energy that would fuel modern life from transportation to medicine, from agriculture to information. But over time, other nuclear imaginaries have emerged, and our ideas about how nuclear energy might transform our lives have changed.^k

In the remainder of this paper, I want to use the concept of sociotechnical imaginaries to construct (and reconstruct) the technological trajectories of three reactor designs and the social orders these trajectories embody: the “Chernobyl-type” RBMK reactor, the NuScale SMR, and the TerraPower fast neutron reactor. I argue that each of these designs came (or comes) with a vision of how the world ought to work and with a set of expectations about how this particular technology would change our world. Furthermore, each of these sociotechnical imaginaries was or will be subject to transformation as actual events unfold (e.g., the Chernobyl disaster, the U.S. trade war with China) and as our perception of what problem we want to solve changes and challenges the very categories of analysis we’ve gotten used to (e.g., nations fearing proliferation or humanity anticipating climate change).^l The sociotechnical imaginaries I associate with these particular technologies and their vision of and for the world have been, and will continue to be, at odds with one another. It is still unclear which one, or which ones, will prevail and how.

IV. DISCUSSION: NEW DESIGNS

Designing new nuclear reactor types seems to come in waves. During the 1950s and 1960s, many nations experimented with different materials and processes, and scientists eagerly exchanged their experiences and shared their experiments with their international peers. In 1955, 1958, 1964, and 1971, the United Nations sponsored four conferences on “Peaceful Uses of Atomic Energy” in Geneva, Switzerland. As the diminishing frequency of these meetings suggests, by the early 1970s most nations had completed their research and development (R&D) programs and focused on one or two designs to develop their respective industries.^m In the decades since, we haven’t seen drastic design innovation; most of the design bureaus’ efforts have been focused on perfecting existing reactor lines.ⁿ With the end of plant life looming, the early 2000s saw another peak in new design development, along with talk of a “nuclear renaissance.”^48,49

Perhaps not surprisingly, most of the “new” designs proposed during this period had actually been around since the initial reactor design period in the 1950s and 1960s. Scientists and engineers back then had the liberty to play with all kinds of ideas, both in the Soviet Union and in the West. The initial Geneva conferences had required significant declassification in advance of the meetings, as well as facilitated targeted espionage under the cloak of scientific diplomacy. The meetings also served as sounding boards for big visions of how nuclear energy was going to change the world.^50,51 At the third conference in 1964, one Soviet scientist, Andranik Petrosyiants, left no doubt about the trajectory of nuclear energy’s world-transforming power: this was just the beginning, he stated. From today’s inefficient slow-neutron fission machines, nuclear energy would march toward fast neutron reactors, and ultimately fusion energy.⁵² Implicit in this vision was that soon, the world would no longer need to burn fossil fuels or dam up rivers, even as more and more aspects of life were being transformed by abundantly available electricity. This technocratic imaginary, of course, seemed even more plausible in a state already dedicated to radical social reform. But what about specific proposals, specific imaginaries that accompanied (or accompany) specific designs—reactor designs that not only promise technical novelty but also new social worlds? Let’s take a look at three examples. I will briefly provide some historical context on each design’s development, and then discuss how the concepts of safety, security, and sociotechnical imaginaries were (or are) framed for each design.

IV.A. RBMK Design

It was 1965 and Soviet economic leaders had to make a decision. The country’s nuclear power program was in trouble—production bottlenecks and delays jeopardized the projected growth of the nuclear industry. The reactor design adopted for future Soviet nuclear plants, the VVER, a pressurized light water reactor (LWR) similar to the PWR designs dominant in the West, had proven difficult to manufacture, with only one Soviet factory capable of delivering pressure vessels that met the high-quality criteria required. After almost two decades of heated debate over the value of developing a nuclear industry at all, in a country with vast natural resources and no imminent energy crisis, nuclear proponents had finally managed to get the allocations for nuclear new build written into the short-term and long-term plans—the backbone of the Soviet economy. Letting these plans slide now would mean the certain end to their nuclear ambitions.⁶

The reactor designers from NIKIET, one of the leading design and construction bureaus at the time, believed they had a solution.^o Based on their vast experience with “industrial reactors,” code for reactors that produced plutonium for nuclear weapons, they proposed a second design for the country’s fledgling nuclear sector: a graphite-moderated, water-cooled BWR, the so-called RBMK, short for reaktor bolshoi moshchnosti kanal’nyi/kipiashchii (high-power channel/BWR). In contrast to the VVER, this model could be assembled mostly onsite and didn’t require a complicated factory-manufactured pressure vessel.⁶ Furthermore, the RBMK relied on well-established supply industries, but on different branches from those supporting the VVER, thus allowing it to be produced with little impact on the VVER line.^7,53,54 And, while the VVERs at the time were built with small power capacities [210 and 365 MW(electric)] and slowly increased in output to 440 MW(electric), the RBMK entered the scene as a 1000 MW(electric) “giant.”^p NIKIET’s engineers argued that they had ample experience with so-called “dual-use” reactors whose main purpose was to produce as much weapon-grade plutonium as possible. These reactors had been modified to provide heat and electricity for nearby towns as byproducts.^q NIKIET claimed that modifying the graphite-water design to entirely civilian machines was within reach, and they offered the RBMK as a supplement to the already approved line of VVERs.

Ultimately, Soviet decision makers approved the RBMK proposal, despite the fact that peer reviews of the original design suggested that significant changes were needed.⁶ The Kurchatov Institute, the leading nuclear science center in Moscow, joined NIKIET in redesigning and implementing the RBMK (Ref. 55). The idea for the RBMK’s graphite-water core is credited to Igor Kurchatov and Savelii Feinberg, a lesser known but highly accomplished physicist at the same research institute. In addition to chairing a reactor design division at the Kurchatov Institute, teaching at Moscow’s finest Engineering University, and authoring scientific patents as well as textbooks, Feinberg excelled at inventing original, if sometimes outlandish, reactor designs and had them built as prototypes at a reactor park near a city then called Melekess (today Dmitrovgrad).^4,56,57,r Feinberg was the second lead designer on the RBMK development, working directly with the Kurchatov Institute’s director, Anatolii Aleksandrov. Feinberg unexpectedly passed away two weeks before he was supposed to chair the first RBMK unit’s startup commission in 1973. One of his closest collaborators reported that just weeks earlier, Feinberg had told him that they would need to return to their calculations and “check and refine” everything.⁵⁹ Feinberg’s successors did not follow through with these plans; the construction of more RBMKs continued and accelerated without the thorough review Feinberg had planned.

Even though from today’s point of view we may recognize clear safety issues with the RBMK design, it is worth noting that at the time, some engineers considered the RBMK design safer than the PWR design. They described its many individual fuel channels as “modular” and argued that an accident would unlikely affect all of them at once.⁶⁰ In a PWR, by contrast, the entire core would be affected by an accident in a single fuel channel. Safety concerns reappeared over the course of the next decade, but RBMK designers also continuously implemented new and enhanced safety features that made every new RBMK safer than the previous one. Over time, emergency cooling systems, accident localization provisions, emergency power supply, and other features were added or improved.^7,61

It was only the Chernobyl disaster in 1986 that eventually led to a fundamental overhaul of the RBMK design. From raising the fuel enrichment to preventing the withdrawal of too many control rods, from drastically improving quality control to requiring simulator training for operators, Chernobyl challenged assumptions about the nuclear safety of RBMKs and beyond. And while the disaster was the result of an unfortunate combination of technical problems, human error, and more systemic issues, it is also important to acknowledge that the same combination of technologies, nuclear professionals, and specific features of the Soviet industrial, economic, and administrative system allowed a fantastically fast (from today’s perspective) expansion of the nuclear industry, an industry that managed to deliver almost 30 GW(electric) of nuclear capacity by 1986 (Ref. 6).

A particularly attractive bonus for decision makers steeped in the Cold War nuclear arms race was the fact that the RBMK was refueled online and could therefore be adjusted to the needs of national security. Your military needed more weapon-grade plutonium? In a pinch, the RBMK could deliver. This argument was mostly rhetorical, however: as the experience with dual-use reactors quickly made clear to Soviet managers, using one and the same facility for secret, military purposes and also for civilian use created more problems than it helped address.^s There is no evidence in the open record that any RBMK was ever used to actually augment the vast Soviet plutonium production capability. Nevertheless, the Soviets did recognize the potential of the RBMK to modify its fuel burnup, and thus its proliferation potential. They never exported a single RBMK abroad, not even to their closest allies.^6,7

At the same time, security in the Soviet Union was quite different from the challenges Russian nuclear facilities face today. The closed character of Soviet society and the lack of a market for nuclear materials made it to some extent irrelevant how well secured nuclear facilities were or what happened with the plutonium in the spent fuel. The RBMK’s online refueling feature was seen primarily as a means of maximizing burnup and thus electricity generation from a given fuel load, which presented another advantage over the VVER design, which has to be shut down for refueling.

The sociotechnical imaginary that sustained the production of over a dozen RBMK reactors, including two with 1500 MW(electric) capacity each in Lithuania, was one that offered Soviet industry and citizens an unlimited supply of electricity, which was seen as the backbone of the people’s economy and would make progress, modernization, and ultimately communism, not just possible but imminent. The RBMK promised to realize the rapid expansion of nuclear energy in the Soviet Union. It built on the idea of ample baseload generation, the use of prefabricated parts that didn’t require sophisticated factory manufacturing, and easy onsite assembly instead of difficult transportation that constrained the reactor’s size parameters.^t Tracing its lineage to military production and dual-use reactors, the RBMK also tapped into the imaginary of “the electrification of the entire country,” a slogan dating back to Lenin’s famous “State Electrification Plan.”^62–65

There were important economic, technical, and political constraints on the preferred reactor design at the time—the VVER—making it clear that it couldn’t carry the planned expansion of the nuclear sector alone. But why, of many other options, including operating ones, the RBMK emerged successfully is not easily explained based on technical or economic factors alone. In the mid-1960s, neither reactor designers nor economic planners could know which reactor type would eventually prove technically and economically superior. At the time, the RBMK “was neither the cheapest nor the technologically most sophisticated design available, nor did it have the most substantial track record.”⁶ But its promoters managed to galvanize support for it by linking its technical features and economic capacity to its promise of social transformation, cultural uniqueness, and strategic potential. By doing so, the RBMK’s promoters actively engaged in crafting a sociotechnical imaginary and firmly anchoring it in the country’s institutions.

Today, the RBMK design has been abandoned for all practical purposes (if only in terms of new build, not of continued operation), yet its imaginary—that of boundless energy and reliable baseload generation—survives. In a way then, the RBMK’s sociotechnical imaginary has done its job, it has helped the Soviet nuclear power program succeed.

IV.B. NuScale’s SMR Design

Small modular reactors are usually defined as “units with a generating capacity of less than 300 MWe” (Ref. 36). SMRs are in the early stages of development, with many design parameters still to be defined.^66–69 One of the main advantages their proponents are emphasizing is the reduced cost, especially in terms of initial capital investment compared to traditional large reactors. SMR developers expect modular designs and construction processes will generate economies of numbers and open up multiple supply opportunities. NuScale, a startup company based in Oregon that had closely worked with Oregon State University (OSU) scientists since the early 2000s, acquired the exclusive rights to a small reactor design through a technology transfer agreement in 2007. The company has estimated its first plant will cost just under $3 billion to build.^70,71

The brainchild of Dr. José N. Reyes, NuScale’s chief technology officer and nuclear engineering professor emeritus at OSU, this 50 MW(electric) pressurized water design uses uranium enrichment levels of under 5% for its standard PWR fuel assemblies and a 24-month refueling cycle. Its small modules are designed to operate underground, and the reactor relies on cooling water that naturally circulates through the core.^71,72 The main difference to conventional PWRs is its size—NuScale’s core has only 1/20 of the nuclear fuel contained in the more typical PWRs operating today—and the “simplification or elimination of systems,” along with improved, partly passive safety features that are intended to prevent the possibility of a meltdown.^73,u NuScale relies on factory-built modular components that are produced and assembled offsite before being shipped or trucked to the site. A standardized manufacturing process, the company argues, “increases efficiency, improves quality, and lowers cost.”⁷⁴ Up to 12 modules can be connected to one another, providing a 600 MW(electric) assembly.⁷¹ Cooperating with multiple international partners, NuScale has built and operates a domestic testing facility that includes an electrically heated core at one third of the actual size to simulate operating conditions and demonstrate the design’s viability.⁷⁵ The company also received a coveted financial assistance award through the U.S. Department of Energy to support bringing this design to market.⁷¹

The NRC began reviewing NuScale’s Design Certification Application in 2017 and completed its technical review in September 2020, with a final report expected in early 2021 (Refs. 71 and 76). NuScale has been the first SMR to undergo this review and could well be the first company to actually build an SMR in the United States. NuScale has been collaborating with Idaho National Laboratory to site its first 12-module SMR, which will be operated by Energy Northwest, a consortium of electricity utilities.⁷¹

NuScale claims that its reactor can’t melt down.⁷¹ In terms of safety, NuScale’s design is clearly taking into account accident conditions in the design of its emergency cooling system that other designs don’t. More than any other SMR design currently being developed, NuScale emphasizes its accident resilience and superior safety features that rely on simplifying and decoupling complex systems.^77,78 One of these safety improvements has to do with eliminating pumps to drive the cooling water through the core. Instead, NuScale relies on natural convection, which eliminates moving parts that might be vulnerable to failure in an accident. Furthermore, NuScale’s reactors will be protected by much smaller containment vessels than regular PWRs, which drastically increases their capacity to withstand pressure.⁷⁹

When my colleague Ross Carper and I talked to Bruce Laundrey, then NuScale’s chief marketing officer, in 2010 in the context of preparing a research paper on SMRs, we learned that NuScale was very consciously designing its reactor to fit within the existing regulatory framework.⁸⁰ Yes, the engineers understood that other designs might offer higher efficiencies, but in order to get their new design expeditiously processed and eventually approved by the NRC, they had decided that it was worth sticking with a LWR design, just an underground, miniaturized, safety-improved version of it. We learned in interviews with other designers that the existing regulatory framework significantly shapes design choices and that reactor designers consciously position themselves and their reactor design either within or outside of these frameworks. Either decision involves consequences in terms of anticipated duration of review by the regulator, depending on the existing expertise to evaluate and eventually approve a given design. These designs, then, are either “evolutionary,” those that consciously build on existing design choices and modify them gently so as to enable the licensing agency to expeditiously review and hopefully approve their proposal, or “revolutionary,” those that propose novel solutions that the regulator is still unfamiliar with, conscious of the risk of delay, but adamant about the value of their proposal.^v

NuScale’s SMR fits squarely into the evolutionary category. With its explicit reliance on tried and true technical components in addition to its novel system development, it taps into current ideas of what constitutes nuclear safety and how safety is assessed, licensed, and monitored. By deliberately flying under the radar of currently accepted licensing guidelines, and by presenting its power module as an evolutionary design based on the proven technologies of the past, NuScale avoids the extra scrutiny that would apply to a more revolutionary design.

In terms of security, some designers have promoted SMRs with “a long-lifetime sealed core [that] could reduce opportunities for material diversion.”³⁶ NuScale is not one of them. Their insignificantly longer-than-average refueling cycle of two years means that every other year a module has to be shut down and its fuel replaced, which also means storing fresh fuel nearby at certain times and securing the spent fuel. On the other hand, NuScale’s fuel enrichment (at just under 4%) conforms to current industry standards and won’t raise any additional proliferation concerns. And even though NuScale’s reactors will produce plutonium in the core, (1) the cores are small and (2) the cores are sealed until maximum fuel burnup is achieved, at which point the isotope ratio in the spent fuel is no longer ideal for the extraction of weapon-grade material.

The sociotechnical imaginary that NuScale offers involves, quite prominently, the idea of “highly reliable nuclear power for mission-critical applications”; in other words, nuclear power for deployment at military bases, in remote locations, or for use in oil and gas development.⁸⁴ Scale has become a negotiable variable again. Instead of simply pursuing economies of scale (the bigger the better/cheaper), more recent developments in grid structure (feed in by small rooftop units, electric cars, etc.), changes in electricity demand, and a more nuanced grasp of demand fluctuation has helped the idea of a small yet modular design gain traction. SMRs are able to link up with this new energy policy context where large centralized generation is no longer the main kind of power needed; load following and grid sharing with many individual feed-in stations has become increasingly important.

In other words, the sociotechnical imaginary of SMRs no longer envisions societies depending on large, centralized baseload stations with smaller peak-load capacities covering early morning and evening surges. Instead, they tie into a vision of the world where individuals generate energy at home and feed it into the grid. This happens not always at predictable times since our lives’ rhythms are also envisioned as much more fluent and idiosyncratic. The SMR’s sociotechnical imaginary plays on the idea that theirs is one among several, and its modular structure can accommodate either or both: SMRs can fulfill small baseload demand or they can be grouped together to form larger units. Depending on specific energy needs, they offer more flexibility in satisfying demand and decentralized supply.

IV.C. TerraPower’s Fast Reactor

Around the same time OSU granted NuScale Power exclusive rights in 2007, TerraPower chose a very different path. Funded by private capital, something unique to the United States, TerraPower clearly falls into the category of revolutionary designs, even though fast neutron reactors (variously referred to as breeders or burners) have been around for decades and were built and operated, albeit not always successfully, in several countries from the 1950s onward.⁸⁵ Fast neutron reactors typically allow for flexible core arrangements that can be configured so as to “breed” more fuel or to “burn” long-lived actinides, depending on the initial fuel composition. Originally viewed as a solution to the scarcity of nuclear fuel, nowadays they promise a solution to the spent nuclear fuel (and more generally waste) problem that haunts the nuclear industry. Historically, fast neutron reactors have not had the best safety record. The combination of materials and complicated systems has repeatedly led to fires and other accidents in existing facilities. Both Japan and France shut down their respective facilities (Monju and Superphénix) after a series of accidents.^w The formerly Soviet reactor in Shevchenko, now Aktau, Kazakhstan, has also been permanently shut down and is undergoing decommissioning.⁸⁷ Currently, only Russia operates commercial-scale fast neutron reactors at the Beloyarsk site in the Ural mountains (BN-600 and BN-800, referring to their megawatt-electric output).

TerraPower’s so-called traveling wave reactor (TWR) features a small core of nuclear fuel (enriched uranium, for example) sited in the center of a larger mass of nonfissile material, such as depleted uranium. Neutrons from the core transform ²³⁸U into ²³⁹Pu, a process referred to as “breeding”:

Over time, enough fuel is bred in the area surrounding the core that it begins to undergo fission as well, sending neutrons further into the mass and continuing the process while the original core burns out. Over a period of decades, the reaction moves from the core of the reactor to the outside, thus giving the name “travelling wave.”⁸⁸

TerraPower’s reactor would supposedly operate without refueling for a period of 40 to 60 years and is intended to be sited below grade, that is, underground. TerraPower claims that their design avoids reprocessing and burns what is now considered waste, providing a much more efficient use of nuclear fuel than current thermal reactors.

The Ur-designer of the TWR, invoked on the company’s website, was Savelii Feinberg, the Soviet nuclear scientist we met earlier.^x According to TerraPower, Feinberg presented the idea of a reactor that would slowly burn through its fuel like a smoldering cigar, breeding its own fuel in the process, in 1958.

TerraPower correctly anticipated trouble building a prototype TWR in the United States. In September 2015, the company signed an agreement with the China National Nuclear Corporation (CNNC) to build a 600 MW(electric) prototype reactor with a construction start in 2018 and a projected completion by 2025. Commercial units were expected to follow suit.⁹⁰ In 2019, however, TerraPower abruptly abandoned these plans as a consequence of the restrictions imposed by the Trump administration, ostensibly “to prevent China’s illegal diversion of U.S. civil nuclear technology for military and other unauthorized purposes” and more broadly to counter repeated instances of intellectual property theft by China related to nuclear energy.^90–92

TerraPower’s design doesn’t lead with talk about improved safety features; the cost and efficiency argument dominates. This is no coincidence. The liquid sodium used to cool the core is a known risk. While more efficient than water at moving heat and less corrosive to metal piping, it is highly toxic and instantly ignites when it encounters oxygen. Furthermore, the company decided to modify the design in 2011 due to concerns about how to cool a moving region (the wave). The modified design still starts the fission reaction in the center, but keeps the “breeding” in the central region by having an overhead crane system gradually and progressively move the outer edge of the core toward the central region while moving used fuel pins to the periphery.⁹³

Fast neutron reactors are flexible, and therefore highly problematic in terms of proliferation because they can be configured either to breed more plutonium or to burn long-living actinides that create the most difficult challenges for spent fuel management. In terms of security and safeguards concerns, the TWR designers argue that the reactor allows the transmutation of materials inside the heavily protected reactor that would otherwise require vulnerable reprocessing facilities. Fast reactors not only produce significantly less waste than current designs, they can burn the radioactive waste produced by conventional LWRs. Still, the fact that these reactors can breed plutonium (including the weapon-grade kind) makes them a higher proliferation risk than average LWRs. Proponents argue that the plutonium the reactor breeds doesn’t last long before it gets split into “a cascade of fission products almost immediately.”⁹³ And yet, like other designs that rely on online refueling, the TWR’s core can theoretically be arranged to produce substantial quantities of weapon-grade materials.

The same flexibility that allows this design to offer new approaches to spent fuel and other radioactive waste treatment (rather than dry-cask storage or deep geological repositories), then, also makes these kinds of reactors an increased proliferation risk. Malicious acts at both state and nonstate levels are not only easily imagined but seem to require intense additional security efforts. For example, before Russia’s second fast reactor at Beloyarsk, the BN-800, became operational, occasional shutdowns of the first one (the BN-600) led to security concerns among the facility’s international partners who sent their spent nuclear fuel there to have it “burned.” Without a backup, the spent nuclear fuel, containing significant amounts of plutonium, in these instances was sitting in a rather unprotected state at the site waiting for the reactor to restart.^y To summarize, fast reactors are more flexible, but also more vulnerable in terms of proliferation than conventional LWRs (Ref. 36).

As mentioned above, in the early days of nuclear power breeder reactors in general were seen as the next step in the development of nuclear energy. The so-called “plutonium economy” they enabled was supposed to alleviate scarce uranium supplies and make the nuclear industry self-reliant and autonomous; it would allow a society to emerge that never again had to worry about pollution or electricity.⁶ The sociotechnical imaginary TerraPower puts forth, most prominently through its spokesman and private investor Bill Gates, is that of a hip twenty-first century technology that will not only fulfill the dream of boundless energy, but will also, once and for all, take care of the nuclear industry’s waste problem. Furthermore, TerraPower portrays its innovative design as part of a new model of society, where private-public partnerships—private companies financing ideas and public funds supporting full-fledged programs—drive progress and innovation. The now defunct agreement with CNNC reveals that growing energy demand and a long-term strategic energy plan may be vital elements of TerraPower’s vision.^90–92 To save the prospects of the TWR design and build a prototype in the United States, TerraPower views both federal support for advanced reactors and significant changes to the NRC’s safety evaluation review as essential.⁹⁴ While some journalists recognize this as quite a plan, considering the potential consequences, such requests are rarely flagged as what they are—not merely business plans, but world-making propositions.^90,z

These three designs, then, differ not just in terms of technology but bring to the table very different ideas about what our future nuclear-powered world ought to look like. Do we worry about safety, and if so, how does each specific design address this concern? How do we make sure that sensitive material doesn’t fall into the wrong hands and that potentially dual-use processes are adequately protected? And finally, given these requirements, how does the world each particular design thus envisions look like: What regulatory infrastructure do we need? What governance structures do we anticipate? Whose preferences do these visions reflect? Each reactor design embodies a vision of a sociotechnical future and seeks to introduce this vision into our existing regulatory and governance framework, whether consciously or not. It’s worth deliberating these visions carefully to ponder what they emphasize and what they leave out, whose ideas they support and which options they curtail, and what sociotechnical preferences they maintain or reinforce.

V. CONCLUSIONS

Contrary to popular belief, and recent visual narratives about Chernobyl, nuclear reactors are never cheap, and the idea that economic (or technical or political) considerations alone would influence the adoption of any given design are misguided at best and ludicrous at worst. With any complex technology, there are a myriad of factors that contribute to decisions about its adoption and wide-spread implementation, exemplified by the extended process it takes to design, prototype, test, and eventually standardize a commercial nuclear reactor design, back in the 1960s as well as today. I have suggested ways of making our implicit frameworks visible, if for no other reason than to actively and consciously tweak them, while staying aware of the implications that technical choices may have on our nontechnical expectations and vice versa. I hope that the conceptual framework I have presented, along with three specific examples, is one that can assist practitioners in making explicit the social worlds they envision when designing tomorrow’s reactors. Viewed through an STS lens, it becomes abundantly clear that it’s never just about the technology: designers make worlds and technologies structure our social experience. Successfully aligning new technologies and visions of our future world will require a broader approach to designing and might involve the inclusion of more than technical voices in defining desirable sociotechnical imaginaries.^aa What social problem is it that a new reactor design attempts to solve? Who gets to decide whether the solution proposed by the designers is credible and feasible? How might a modified design address central issues more effectively? A more broadly imagined sociotechnical future might reveal important insights, both in terms of limits and opportunities, for the nuclear industry.

Let me briefly recap how the three examples I discussed speak to this point. When the RBMK design was adopted as the second type for the Soviet nuclear fleet in the mid-1960s, it promised not only fast, ample generating capability based on a tried and true design template, it also fit the requirement of the time to offer a backup capability for plutonium production due to its online refueling mechanism. A feature like this would make the RBMK design less attractive in today’s context where proliferation resistance is valued, but at the time, the Cold War was at its peak and this option made the design appear more attractive, not less, to decision makers.

What we’re facing today in terms of new designs is strikingly different. The need for large, centralized baseload generating capacity has somewhat diminished, enabling the idea of SMRs to gain traction. What is more, the excess production of weapon-grade material is no longer viewed as an asset, but rather framed as a serious proliferation risk, particularly for export, and therefore to be avoided at all cost. Safety has also taken on a new significance in the aftermath of several severe accidents at nuclear power reactors. The recognition of nonnegotiable safety requirements has prompted various attempts to “build in” safety, from automating more systems and relying less on human expertise and judgment, and from making high-quality simulator training mandatory to relying on “passive safety” systems to offset the need to operate any additional equipment in crisis situations.

All of these initiatives entail trade-offs; we should carefully weigh these options. The multitude of proposals that have been put forward in recent years makes it more difficult to navigate security and nonproliferation concerns as we try to evaluate them in terms of how they fit into our society and into what we hope our future societies might look like. The diversity of designs may help the goal of inherent safety and proliferation resistance and also meet some other goals (such as reducing the amount of waste produced), but it raises new problems, for example, for emergency preparedness and response, as it may require a number of response strategies instead of one. Building in capabilities to assist a system may be at odds with sealing it off from outside interference. The safety/security tensions are inescapable yet often avoided when it comes to new designs, either in scale or in substance. Multinational fuel cycle facilities may turn out to be the most effective institutional barrier to proliferation—more than technical and material criteria taken together—but they may face tough legal and political obstacles. Other factors to be considered might include the number and training of operators at reactors and fuel cycle facilities, a state’s industrial capacity and experience managing different types of nuclear facilities, national commitments to peaceful uses of nuclear energy, regional agreements, and technical cooperation, to name but a few.³⁶

Today’s designs in the United States are being developed by entrepreneurs rather than national laboratories or other government organizations, diminishing the role of the state and increasing instead the burden on existing regulatory institutions that were set up to license and regulate a limited set of designs. The mutual expectations of regulators and designers continues to shape what startup companies and inventors put forth, almost regardless of the purported efficiency of a new design. Designers of revolutionary reactor ideas may complain about constraining regulation, but in a sense, it’s almost like self-censorship: The gate-keepers to new designs don’t know how to evaluate new designs, therefore the designers either shape their designs to fit within the existing regulatory structure or they attempt to avoid that regulatory structure by moving abroad for prototyping.^bb On the other hand, one lesson we can learn from the exclusively state-funded, centrally managed industrial system of the former Soviet Union is that while it enabled top-down decisions and quick implementation, it held up regulation, supervision, and quality control.

I have argued elsewhere in more detail that the RBMK, even in its early, rough design that first materialized near what was then the beautiful city of Leningrad (today St. Petersburg), made sense at the time. It fit into an imaginary of technological optimism, of Soviet modernity, and of scientific and engineering prowess.⁶ It made the hard-won plan to expand a domestic Soviet nuclear industry possible, and quickly. The RBMK was born into a system of scarcity and ambition, ingenuity and fatalism, secrecy and propaganda. That’s where it both succeeded and failed. With our new designs, one apparently inspired by the same scientist as the RBMK, we have a chance to consider what worlds they suggest, what institutions they’d require, which discourses they’d facilitate, and whether we’d rather question, revise, and adjust those and other ordering instruments to the world we’d like to live in.⁹⁹

At the beginning of this paper I argued that nuclear reactor design is never simply a matter of technical functionality. Whether a design is good or bad, functional or faulty, successful or not, is not an immediate outcome of technical calculations. What made the RBMK succeed? Why is the TWR struggling, despite its impressive pedigree? How can one man design reactors that inspire different worlds and different times? We need to broaden our view of reactors to include the sociotechnical visions they might entail: What kind of society relies on large, centralized machines as opposed to small distributed ones? What kind of world would an efficient, flexible, yet vulnerable fast neutron reactor support? How could we craft a social order that is capable of combating climate change and getting rid of nuclear weapons? A society that embraces decentralized autonomy (and energy generation) and yet unites behind one (or several, ideally compatible) sociotechnical imaginaries? And who gets to make these choices: Our engineers? Our economists? Or our billionaires? Contrary to Vernadskii’s optimistic prediction from 1922, our options in 2020 are no less fraught, consequential, and exciting than when we first started introducing nuclear reactors into our world.

Acknowledgments

This work was supported by the National Science Foundation [SES1351575]. I am grateful for the feedback on earlier versions of this paper that I received at the conference, “The Nuclear and Social Science Nexus,” at the Nuclear Energy Agency in Paris in December 2019, and especially for the detailed comments and helpful suggestions from Taylor Loy and three anonymous reviewers, some of whom identified as nuclear engineers. All remaining errors are mine.

Notes

^a VVER stands for vodo-vodianoi energeticheskii reactor, water [cooled], water [moderated] power reactor.

^b The fact that we haven’t seen this sustained production of reactors in many countries may be one reason why the nuclear industry system is so strained today.

^c Hughes coined the term “large technological systems” to describe highly complex systems that contain not only technical artifacts, but also social components such as organizations, scientific knowledge, legislative artifacts, and regulatory structures.

^d While this type of technological optimism was not limited to the 1960s, it was that decade, wedged between post-war optimism and the period when the nuclear “bandwagon” got going, that shaped a more specific kind of “nuclear optimism” that resulted in at times rather utopian ideas about the transformative potential of nuclear energy.

^e To name but a few of such initiatives, Virginia Tech recently launched an interdisciplinary graduate certificate in Nuclear Science, Technology, and Policy, the University of California Berkeley has successfully connected its History of Science and Nuclear Engineering programs, and Aditi Verma benefited from collaborations among nuclear engineers and STS scholars during her studies at the Massachusetts Institute of Technology. Finally, the workshop this special issue resulted from involved a scientific committee composed of humanities, social sciences, and engineering researchers.

^f The interviews were semistructured and questions covered the history of the designs, technical details, regulatory milestones, plans, and prognoses. They complemented the analysis of regulatory documents and media analyses and answered questions arising from other data. I conducted the interviews on the RBMK alone and in Russian; for the others, I was joined by my colleague Ross Carper.

^g Very specific definitions of “inherent safety” were developed much later.²⁸

^h In addition to the U.S. Nuclear Regulatory Commission (NRC), such regulatory infrastructures include Ukraine’s State Nuclear Regulatory Inspectorate, Russia’s Rostechnadzor (which combines nuclear with environmental and industrial regulation), and European Union organizations such as the Western European Nuclear Regulators’ Association (WENRA) and the European Nuclear Safety Regulators Group (ENSREG) that coordinate European regulators, along with guidelines developed by the International Atomic Energy Agency (IAEA) (see, e.g., https://www.iaea.org/topics/regulatory-infrastructure and http://www.ensreg.eu/members-glance/national-regulators.

ⁱ I am aware that this simplification doesn’t do justice to the nuances of “automation.” And while the rise of human factors research, a post-Chernobyl emphasis on “safety culture,” and various systems of peer review [exemplified by the Institute of Nuclear Power Operations (INPO), the World Association of Nuclear Operators (WANO), and the IAEA] could be interpreted as increased awareness of the nontechnical aspects of safety, they often inadvertently substantiate the idea that technology left to its own devices (that is, free of human intervention) would be safer. Despite the undeniable progress made with regard to nuclear safety over the past decade or two, we tend to see a return to technocratic ways of thinking whenever something goes seriously wrong, the Fukushima disaster being the most recent example. The resistance to relying on human judgment, and on training human judgment to handle unprecedented situations competently and creatively, is alive and well.²⁹

^j The point here is less that current decommissioning practices fail to address specifically below-grade reactors, but rather that end-of-life management of nuclear facilities in general has had a troubled history.

^k Such alternative imaginaries include outright negative ones, as well as others that envision nuclear energy’s role in mitigating global climate change.

^l Recall Bijker et al.’s discussion of the bicycle mentioned previously. Was the bicycle the solution to the problem of transportation, of leisure, or of women’s emancipation? The perception of what exactly was the problem shaped the design of the bicycle itself, which in turn shaped the world it eventually helped make.⁸

^m For an overview of the Oak Ridge National Laboratory’s reactor R&D program, which may have been comparable to the long-lived Soviet one, see Ref. 44, and for France, see Ref. 3.

ⁿ Arguably, there was another peak in activity during the late 1980s and early 1990s when the designs currently under construction (the AP-1000 and the EPR-1000) as well as the ESBWR and the ABWR were developed.^45–47

^o NIKIET stands for Nauchno-issledovatel’skii konstruktorskii institut energotekhniki (Scientific Research and Design Institute of Energy Technologies). Created in 1952, it is now named after its first director Nikolai A. Dollezhal’.

^p The first VVER-1000 came online only in 1980, some seven years after the first RBMK.

^q One of the famous ones was the “Second Ivan” (EI-2). The Soviet delegation at the second Geneva conference reported on the startup of this “Siberian nuclear power plant” in 1958.

^r Melekess/Dmitrovgrad is home to a nuclear research institute, and over the years, eight reactors of various designs were built there.⁵⁸

^s This was primarily due to conflicting secrecy regimes at military and civilian facilities. Thus, the goals of maximizing electricity production on the one hand and weapon-grade materials on the other were quickly found to be incompatible.

^t The VVER’s pressure vessel could only be made at one specialized forge in the USSR, the Izhora Works. The vessel had to fit on a standard rail car or barge and was further constrained by the width of bridges and tunnels it had to be transported through on its way to its destination.

^u The idea of simplifying or eliminating systems has been well established in nuclear reactor design practices since the 1980s (Refs. 45 and 46).

^v This juxtaposition between evolutionary and revolutionary has been used since the 1980s (Refs. 28, 81, 82, and 83).

^w India and Japan have test facilities.⁸⁶

^x “Engineers and researchers have long dreamed of a self-fueling source of energy. As early as 1958, Saveli Feinberg imagined a nuclear reactor that could breed fuel within its core. TerraPower will use proven fast reactor technology, high-performance computing simulations, and real testing in current fast reactor test facilities to make the TWR concept a reality.”⁸⁹

^y Personal conversation by the author with a French representative during a conference in Obninsk, Russia, in about 2006.

^z TerraPower shares these ambitions with NuScale, whose representatives have also demanded significant changes to the U.S. regulatory framework, including the size of the emergency evacuation zone.

^aa There is a vast and growing body of literature in STS that explores different participatory models in technology design and assessment. For recent examples, see Refs. 95, 96, and 97.

^bb Oklo, who recently submitted a license application for their 1.5 MW(electric) Aurora microreactor, is the exception to this rule.⁹⁸