Transnational nuclear knowledge-making beyond science: The laboratory of Chooz A (1960s–80s)

Introduction

Both France and Belgium are among the most nuclear-reliant countries in the world. France derives around 70% of its electricity from nuclear power. ¹ In Belgium, that number is around 40%. ² Moreover, the ties between French and Belgian nuclear engineers and companies are strong. This tight connection has a history – a history that brings us to one specific place in a curious bulge of the French–Belgian border, near the village of Chooz. This article traces the history of the Chooz A nuclear power plant, located in a rock in the border region of the Ardennes. We argue that this reactor was an unusual project that acted for decades as a transnational laboratory in the broadest sense of the word. An incredibly varied group of organizations and individuals from different countries engaged in nuclear diplomacy there, by collaborating and testing out technologies and scientific principles. Yet, it had the specific aim of immediately materializing this knowledge for their own, national projects (Figures 1 and 2).

OPEN IN VIEWER

Figure 1.

The Chooz A nuclear reactor is located in a border bulge on French territory, but surrounded by the Belgian border.

OPEN IN VIEWER

Figure 2.

The Chooz A reactor by the Meuse River in 2014, before its decommissioning. The reactor itself is housed in the hill right behind the large white building (which houses the turbine hall).

Nuclear engineers in France mostly know Chooz A as a decommissioning project, the first of its kind. In Belgium, Chooz A is almost completely forgotten. And despite a renewal of historical scholarship on nuclear energy in France since the 2011 Fukushima accident in science and technology studies (STS), Chooz A has not been the subject of any dedicated research and remains largely overlooked. ³ Yet, we argue that the case of Chooz A teaches us that we need to expand current conceptions of knowledge-making in history and STS. Scholars in those fields have increasingly paid attention to transnational nuclear knowledge-making and the role that national interests and diplomacy play in that. However, ideas about where knowledge about nuclear things is created mostly revolve around laboratories in the narrowest sense of the word, where nuclear scientists do scientific experiments. This article shifts the focus more broadly to all places that acted as “laboratories.” This also includes, in our view, nuclear power plants. We challenge specifically the dominant analytical separation in historiography between scientific knowledge production about nuclear things and nuclear energy infrastructure. Instead, we emphasize the importance of combined local, spatial, and transnational processes of knowledge-making through tinkering and technology in use.

We do this by relying on previously unexplored archival collections from Électricité de France (EDF), the French Nuclear Safety and Radiation Protection Authority (ASNR), Westinghouse, and the Belgian National Archives. To account for the gaps in the archival material as well as to understand the experiences of designing and operating Chooz, we also rely on ten interviews with former personnel. In the first part of this article, we reflect more conceptually on the notions of laboratory and (trans) nationalism and discuss how the case of Chooz A can inform the broader literature in history and STS. The second part sketches how French and Belgian scientists and engineers began engaging with nuclear technologies and collaborating on the Chooz project. The third part focuses on the design process, preparations, and construction of the plant in the mid-1960s. While the design depended heavily on American knowledge, a range of very unique features served to test out new design practices. The fourth part of the article zooms in on the operational problems that emerged at the end of the 1960s, which prompted a crisis of expertise and exposed the laboratory nature of the project. In parts five and six, we demonstrate two other important learning moments at Chooz: the unanticipated contamination of the river with tritium and the introduction of MOX (mixed-oxide) fuel. Lastly, we conclude that Chooz A was a laboratory and demonstrate the different ways in which the project shaped later nuclear development in France, Belgium, and other parts of Europe.

A transnational laboratory

A laboratory – the word originating from Latin meaning “a place to work” – is today conventionally defined as a room or building with scientific equipment for doing scientific tests or for teaching science. But what if it can be more than that? ⁴ STS scholarship has demonstrated that a laboratory does not necessarily need to be located in a science department with scientists in white gowns. It is a place where people experiment and learn. Helen Tilley labeled nineteenth-century Africa a “living laboratory” where British scientists acquired new expertise about the environment, medicine, and people. ⁵ In a broader sense, a laboratory is not just a place for research or teaching. In Laboratory Life, Bruno Latour and Steve Woolgar demonstrated how laboratories are sites where scientific facts are socially constructed and shaped by cultural beliefs and practices. ⁶ The notion of the laboratory is particularly useful because it stresses the relationship between humans and technology, the practical work, locality, and playful creativity. In this sense, it is unsurprising that many university departments are also called “labs” today.

Laboratories or sites of nuclear knowledge-making are a dominant theme in the literature on nuclear science and technology in STS. But energy infrastructures are rarely considered as laboratories. An exception to this is perhaps scholarship on “prototypes” or “demonstrations.” Well-documented cases include the Clinch River reactor, ⁷ Shippingport, ⁸ or Superphénix. ⁹ While a prototype is generally designed to prove feasibility in view of a subsequent series, experimentation within energy infrastructures is often broader. It encompasses the earlier-mentioned aspects of knowledge production, practical work, and creativity. Moreover, not all power plants that were laboratories were prototypes or demonstrations. Chooz A, for instance, was conceived as a full-scale power reactor, the latest in an already established line, with no direct follow-on planned at the time of design or early operation. Previous conceptions of a “laboratory” within STS can thus be helpful to analyze this.

The role of knowledge in nuclear history is primarily interpreted through the lens of “big science,” particularly nuclear physics or chemistry. ¹⁰ An analytical separation in historiography persists between nuclear knowledge-making and nuclear power. Nuclear historians have often even created this separation through establishing causalities, for instance, by demonstrating how scientific insights from laboratories, particularly those working with physics and chemistry, feed right into nuclear engineering and power production. ¹¹ But this is not the full story. In recent years, historians of science and technology more broadly have blurred the distinction between technology and knowledge, pointing to “know-how” and more tacit, practical, and procedural forms of knowledge. ¹² In this context, historians like David Edgerton, Ruth Oldenziel, and Mikael Hård have emphasized the importance of tinkering and using, as opposed to scientific innovation, for creating technological change and knowledge production. ¹³ Although the histories of these scientific discoveries are important, we still lack scholarship that analyzes the transnational ways in which nuclear knowledge is created outside of scientific institutions by experimenting with a technology in more practical ways, through usage, tinkering, and testing.

Crucial to laboratories in this sense, we argue, is diplomacy – diplomacy between national actors, but more often than not transnational actors as well. In the past two decades, scholars in the nuclear humanities and social sciences have particularly focused on transnational exchanges as a reaction to predominantly national frameworks of analysis. ¹⁴ Kenji Ito and Maria Rentetzi have introduced the notion of nuclear diplomacy, emphasizing the role of diplomatic negotiations and national interests in nuclear knowledge-making. ¹⁵ Anna Åberg, for example, has shown how the International Thermonuclear Experimental Reactor (ITER) project in the South of France, the world’s largest nuclear fusion project, was the result of compromise, reciprocity, and complex technoscientific diplomacy. ¹⁶ International organizations were important venues for nuclear diplomacy, too, particularly the IAEA (International Atomic Energy Agency). ¹⁷

An often-underestimated organization in the study of nuclear diplomacy is Euratom. The historiography on Euratom almost unanimously labels the international organization as a failure. According to historian John Krige, Euratom never lived up to its expectations. ¹⁸ Andrew Barry and William Waters observed that the Euratom Treaty was never even fully implemented and was thus an “evident failure.” ¹⁹ However, we will show that Euratom, despite not achieving its initial goals and ambitions, did have a profound influence on nuclear development and knowledge production in Europe. It achieved this through soft power and diplomacy in specific nuclear projects – both through financing and coordinating expertise.

We argue that the Chooz A reactor is a good starting point to begin studying this in more depth. Just like the historiography on the Belgian and French nuclear industries in general, the historiography on Chooz has been dominated by engineering histories and publications from the nuclear industry itself. ²⁰ A group of environmental historians recently studied transboundary connections around nuclear facilities, but the case of Chooz was not part of that project. ²¹ Yet, the experimentation, testing, and tinkering at Chooz proved to have far-reaching consequences for the industry, particularly for building light water reactors (LWRs), today the most prevalent reactor type in the world.

The French and Belgian nuclear programs

The French government turned to nuclear power in the 1950s, viewing it as a promising and economically viable solution for electricity generation. Following the Second World War, national independence became a key pillar of France’s energy policy. Lacking domestic energy resources but high on industrial ambitions, this strategic orientation was institutionalized through the five-year plans (plans quinquennaux). ²² The Commissariat à l’énergie atomique (CEA) initially developed experimental heavy-water reactors in the Paris suburbs (Fontenay-aux-Roses and Saclay) as well as numerous research reactors fueled with enriched uranium supplied by the United States. Although the design of enriched-uranium reactors was clearly inspired by American models, the United States did not assist in the construction of these reactors. The CEA’s first fully operational industrial site was established at Marcoule, in the Gard department, in the early 1960s. The three “G” reactors there produced the plutonium required for the French atomic bomb, the first test of which took place in the Algerian Sahara in 1960. These reactors belonged to the “natural uranium–graphite–gas” (UNGG) line, which formed the basis of a civil nuclear program, with Électricité de France (EDF), France’s nationalized utility, operating reactors at Chinon and Saint-Laurent-des-Eaux.

This dual-use (civil and military) technology was of major strategic interest to France. The use of natural uranium, sourced from mines in both metropolitan France and its colonies, obviated the need for enrichment facilities and thus avoided dependence on the United States. During the 1960s, a gradual transition took place, which historian Gabrielle Hecht has described as a shift from a “nationalist” technopolitical regime, driven by the CEA and its advocacy for French-designed technologies, to a “nationalized” regime led by EDF, whose approach to the nation was grounded in economic rationality and cost-efficiency. ²³ However, critics of the UNGG line emerged and advocated instead for a transition to the American-designed LWR technology, which they deemed more technically mature and economically advantageous. After several years of debate, commonly referred to as the “war of the reactor lines” (guerre des filières), the French UNGG line was officially abandoned in 1969. ²⁴ At that point, however, France had very limited experience with LWRs, and more specifically with the pressurized water reactor (PWR) line. Nonetheless, it could draw upon what, in the 1960s, had been regarded as something of an anomaly or a “five-legged sheep”: the Franco-Belgian Chooz A reactor. ²⁵ A company that had remained relatively inconspicuous until then now entered the picture: Framatome. Established in 1958 by a consortium of companies from the Schneider, Merlin Gerin, and Westinghouse Electric Company groups, the company’s primary objective was to operate under the Westinghouse license to build PWRs.

The American light water technologies EDF was interested in had been part of Belgian nuclear development for a long time. During World War II, the Belgian mining company Union Minière had provided natural uranium, mined in Belgian Congo, for the Manhattan Project in the United States. Through agreements with the United States and the United Kingdom, Belgium gained privileged access to scientific and technological know-how. ²⁶ This led in 1951 to the creation of the research center STK (Studiecentrum voor de Toepassingen van de Kernenergie), later SCK-CEN. ²⁷ At this research institute, Westinghouse licensed its first pressurized reactor outside of the United States: BR3, finished in 1962. ²⁸ By building BR3, the Belgian company Ateliers de constructions éléctriques de Charleroi (ACEC), partially a subsidiary of Westinghouse, acquired expertise in building LWRs, and a generation of Belgian technicians learned how to operate them. ²⁹

Westinghouse’s engagement in the European market was made possible by the U.S.–Euratom agreement of 1958. Euratom had been established to create more political integration in Europe through nuclear development, which was very much seen as a science and technology of the future, in the same way as the European Coal and Steel Community had prompted integration in those sectors years earlier. ³⁰ Euratom was also a crucial part of U.S. foreign policy and President Eisenhower’s Atoms for Peace Program, which aimed to promote the peaceful use of nuclear technologies (as opposed to nuclear weapons production), while at the same time binding its European allies together and securing American dominance in nuclear science and engineering. ³¹ In 1958, the United States and Euratom signed an international agreement that committed the United States to deliver enriched uranium to the Euratom countries, including France and Belgium, while Euratom would facilitate the construction of large-scale American-designed nuclear power plants. The United States and Euratom would also collaborate on research and development in a large Joint Program, allowing significant funds to be invested in nuclear research and infrastructure. ³² But Euratom was more than just research. Article 37 of the Euratom Treaty provides Member States with a right of scrutiny over neighboring countries’ installations, particularly with regard to releases of radioactive effluents. An expert group appointed under Euratom is tasked with delivering an opinion prior to the commissioning of each new nuclear installation in Member States. This gave Euratom (limited) regulatory power as well. ³³

France and Belgium had very different attitudes toward Euratom. The Belgian government, along with the country’s nuclear scientists and engineers, was very enthusiastic. Internationalism, rather than independence, was a key driver of nuclear development for the small country seeking to exercise influence in the postwar period. The French government had a very different attitude. President Charles De Gaulle became notoriously opposed to Euratom. For De Gaulle and his political allies, assisting Euratom would further the interests of the United States and West Germany while compromising French independence. The French government preferred to move ahead alone with its own civil and military gas–graphite program. ³⁴ France only complied minimally with Article 37, seeking to prevent Euratom experts from extending their attention to the safety of French reactors. ³⁵ Still, the CEA and EDF partnered up with Euratom for several fringe projects, including nuclear fusion, breeders, and – at the time – LWRs from the United States.

Chooz A was thus a product of Euratom and the U.S.–Euratom Treaty. Yet, the origins of this French–Belgian–American–Euratom collaboration preceded the Treaty. The idea originated during a visit of European engineers to the United States in 1957, where both French and Belgian engineers toured the newest nuclear reactors. ³⁶ The initiative came from the director of a major Belgian utility, Pierre Gosselin, who contacted EDF after the visit. Building a power plant with Belgian companies could enable EDF to free itself from the oversight of the CEA, which showed little willingness to explore technological pathways outside its own, particularly when they originated abroad. ³⁷

However, Belgian utility companies and EDF had to face skeptical national governments and manufacturing companies. By this time, Charles De Gaulle had become the president of France again. The new government did not look favorably at international nuclear collaborations. Belgium’s pro-European Minister for Economic Affairs, Jean Rey, as well as Euratom, had to apply pressure. In November 1958, they agreed, on the condition that the plant would be built on French territory. ³⁸ For the Belgian utilities, gaining experience from operating a plant that was not located in Belgium was the dream scenario. ³⁹ But some Belgian politicians and manufacturing companies, especially in Wallonia, were skeptical of this. Investing in infrastructure located in France, in a time when the coal and steel industry in Wallonia was in decline, was politically sensitive. ⁴⁰ Meanwhile, manufacturing companies, which had been competing with French companies for years, feared French dominance. The compromise, devised by Gosselin, was to build a second French–Belgian reactor in Wallonia at a later date. This reactor would become Tihange I. ⁴¹ Following the agreement, EDF and the Belgian utilities together created a new common utility: SENA (Société d'énergie nucléaire franco-belge des Ardennes).

Experimenting with PWRs: Designing Chooz

From the outset, it was clear that the French–Belgian reactor would be a unique project – not only in Europe but also compared to reactor projects in the United States. Its most unique design feature was its location: the reactor would be built in a cavern in a rock, located in the hilly Ardennes region. The sources are ambivalent about why it was decided to build the reactor underground. Yet, what is certain is that French and Belgian engineers made the decision as early as 1958, possibly even 1957, so before the project received approval or began planning. One explanation is security. In the context of the Cold War, its hidden location would protect the reactor from attacks or plane crashes. The site of the reactor is close to multiple national borders and the Belgian airbase of Florennes. ⁴² Another explanation is cost. American-designed reactors were usually protected by a containment building, which would contain radioactive particles in the event of an accident. However, these buildings were very expensive. In the middle of the 1960s, engineers in the United States became interested in the prospect of siting reactors underground. They were inspired by examples in Europe, such as the R1 research reactor in Stockholm, Sweden, and the Lucens reactor in Switzerland – both located in a cavern. Yet, these were small research reactors, not a scaled-up commercial nuclear power plant. ⁴³ Even though few people really knew why, there was a clear determination to experiment with a cavern design (Figure 3 and Figure 4). ⁴⁴

OPEN IN VIEWER

Figure 3.

Visualization of the (largely underground) Chooz A reactor.

OPEN IN VIEWER

Figure 4.

Construction work on the cavern.

Soon enough, however, the cavern design presented significant construction and safety problems. One was the quality and strength of the rock. With no in-house expertise on geology, in 1960 SENA contracted several French and Belgian civil engineering bureaus, including a laboratory from EDF, to take geological samples and identify any water leaks. ⁴⁵ Civil engineers identified quite a few leaks in the rock. Moreover, as the caverns were dug, landslides of rocks threatened to hit workers working at the auxiliary building located at the edge of the hill, near the river. Planners had to move the building 35 m, and roofs were built by the entrance to protect from falling rocks. The problems with the rock stability even necessitated design changes to the cavern and eventually delayed the startup of the plant. ⁴⁶

Perhaps the biggest uncertainty was whether the cavern was completely leak-tight. In the reactor containment, whether above ground or underground, the pressure needs to be maintained constantly, which requires the inside to be completely sealed off from the outside. By 1962, it had become clear to the civil engineers that a cavern reinforced by concrete walls and ceilings would not be leak-tight enough. There would be a high risk of water leaking through the concrete into the reactor building. An added risk was that a nuclear accident inside the cavern could create such high pressure that the rock could deform. ⁴⁷ What followed was a series of experiments and tests, taking about a year, to study the water in the rock and construction materials. Engineers proposed to add a thin metal sheet on the outside, between the concrete and the rock, right where they predicted the water leaks would be. Yet, this did not prove safe enough either. SENA eventually decided to surround the entire concrete containment with a steel layer. ⁴⁸ For these changes, it had to actively seek geological expertise abroad. It negotiated with Euratom to receive research funds in the framework of the U.S.–Euratom Treaty. French and Belgian engineers traveled to Pittsburgh to investigate the problem with engineers from Westinghouse and civil engineering consultants. SENA even relied on the expertise of Swedish engineers, who were experienced with using plastics to reinforce and tighten concrete. ⁴⁹

In October of 1960, SENA decided, after considering multiple other designs, upon a PWR type, manufactured by a group of companies connected to Westinghouse – including the Belgian companies Ateliers de Charleroi and Cockerill-Ougrée, and the newly founded French company Framatome. ⁵⁰ For Westinghouse, Chooz was a unique, experimental project and a milestone. ⁵¹ For the French nuclear industry, it was the first time that an industrial reactor with a non-French design would be built on French territory, and together with another country. It was also a reactor without any military purpose, being designed solely for electricity generation. To the Belgians, the BR-3 reactor project, although still under construction in 1958, had provided crucial expertise on PWR design – probably unrivaled anywhere outside of the United States at this point. ⁵²

The design of the Chooz reactor core was equally marked by Westinghouse’s desire to experiment. Westinghouse had primarily developed the pressurized type through the nuclear navy program in the United States. By 1958, Westinghouse had also constructed multiple other full-scale nuclear power plants in the United States. ⁵³ Particularly its latest project had acted as a laboratory: Yankee Rowe in Massachusetts. And it would now also act as a model for Chooz. For instance, both at Yankee and Chooz, the rods that held the uranium fuel were made from stainless steel, not zirconium, the custom material for those pellets at the time. The control rods were also in the shape of a cross instead of a cylinder. ⁵⁴ But translating this design to the French–Belgian context proved to be a challenge. The pressure vessel was an example of that. In a PWR, the core is contained by a large steel pressure vessel. The steel mill of Le Creusot (in Bourgogne–Franche-Comté in eastern France) was tasked with building this vessel. France lacked prior experience in designing steel pressure vessels for this type of reactor. Moreover, French engineers had had bad experiences with using steel vessels to contain a reactor core. In 1959, when EDF opted for a steel pressure vessel for its reactor EDF-1 at Chinon, a crack several meters long developed during construction. This incident represented a serious setback for EDF engineers as they sought to break away from the techniques previously employed by the CEA. ⁵⁵ Le Creusot was thus very dependent on American metallurgical expertise, particularly from the Atomic Energy Commission (AEC), the U.S. Navy, and the American Society of Mechanical Engineers (ASME).

Moreover, Chooz acted as a laboratory for finding solutions to the problem of irradiation embrittlement – a major uncertainty during the 1960s and 1970s. By the time SENA placed the order for a large steel nuclear pressure vessel with Le Creusot, research was beginning to show that steel was more vulnerable to rupturing when exposed to radiation. Scientists were therefore worried that a cool-down of the water in the reactor could potentially cause a vessel break. Initially, Le Creusot designed the vessel according to a code from the U.S. Navy. Starting in the 1960s, the ASME also published new design requirements that were subsequently followed by Le Creusot. ⁵⁶ However, engineers at Framatome and Le Creusot did not wholeheartedly accept the ASME Boiler and Pressure Vessel Code. They found it too detailed. Eventually, a compromise was found, in which Le Creusot was allowed to deviate from the code for the dimensions of the vessel, calculation hypotheses, and calculated stresses. Westinghouse and the ASME would have to agree on the changes, taking over a year and involving American inspections in the French factories. ⁵⁷ The uncertainty regarding the risk of embrittlement rendered American expertise controversial yet indispensable.

Westinghouse and Framatome added a number of additional safety components to the core in the hope of reducing the irradiation and acquiring more insight into the phenomenon. One of them was a thermal screen, a stainless-steel plate attached with bolts between the uranium fuel and the vessel that served to block the irradiation from hitting the vessel. Another solution was the system for neutron flux measurement or “aeroball,” which moves steel balls through the reactor to measure and map neutron flux. ⁵⁸ The first reactor to use this system was Trino in Italy, but the system was tested further in Chooz. Yet, it performed poorly, as the balls frequently became blocked. ⁵⁹ Pierre Schmitt, a former engineer at EDF’s Research and Development Division, recalls that maintenance personnel were regularly called upon to “unclog” the system when the spheres got stuck. ⁶⁰ While French industry abandoned the system, it was recently reintroduced in the Franco-German European Pressurized Reactor (EPR) project.

The design of Chooz A was thus the result of extensive science diplomacy and negotiations between Westinghouse, Framatome, ACEC, Cockerill, EDF, the Belgian utilities, and many other companies and research institutes. But to get a license to operate on French territory, they also needed to negotiate with even more actors. To study the safety of the reactor, a new “joint committee” was founded that united experts from the CEA, Euratom, and the AEC; but also from Italy and West Germany. ⁶¹ In the end, however, a license was quickly granted, and Chooz A went online in 1967.

A French–Belgian social laboratory

The startup of the Chooz A reactor and its first months of operation were a special experience for engineers, technicians, and workers. For both the Belgian and French teams, it was their first time operating a water-cooled nuclear power plant in a uniquely international workplace set in a pristine landscape. Wild boar, numerous in the Ardennes forest, soon became the plant’s emblem (Figure 5). Early staff recall a warm atmosphere and camaraderie; management posts were split fifty-fifty, and conflicts were rare. ⁶²

OPEN IN VIEWER

Figure 5.

The “Sanglier des Ardennes,” here shown on the cover of a report of SENA, was the plant’s emblem.

At the same time, there were notable differences between the French and the Belgians, and nuclear diplomacy was also apparent in the social fabric and daily interactions at the plant. Jacques Daumas, who worked as an engineer at EDF and was sent to Chooz, remembers how his Belgian colleagues all knew each other already. They were very experienced with the reactor type, as they had all been working at the Westinghouse-designed BR-3 reactor in Mol. This had an impact on the language used in the plant. Although the working language was French, the Belgians introduced English jargon, which was otherwise rarely used in France. The Belgians also had an “ultraliberal” mindset, according to Jacques Daumas. ⁶³ They were used to working in private companies. They addressed each other by their first names and were used to negotiating things informally, and often very tenaciously, with their superiors. ⁶⁴ In this way, Chooz could not escape the traditions and cultures of a region strongly shaped by the coal mines and trade unions. ⁶⁵

Meanwhile, at EDF, often described as a state within a state, everything was more formal and regulated. Daumas also recalls how the Belgian engineers and managers were more skilled and practically oriented. In case the operators went on strike, they were trained to take over the operation of the plant. ⁶⁶ Many of the French staff at EDF, on the other hand, came from other coal or gas power plants, and did not know each other. ⁶⁷ And just like the engineers at Framatome, they were a lot younger and inexperienced – for them particularly, Chooz was a learning space. ⁶⁸ That is why a group of five French engineers also went for training in the United States. ⁶⁹ Before being assigned to the Chooz reactor, Jacques Burtheret, for example, had spent a year in Pittsburgh and Saxton, in Pennsylvania, to learn about pressurized reactor operation and safety. ⁷⁰ When the end of the gas–graphite reactor program in France was announced in favor of PWRs at the end of the 1960s, the staff at Chooz were emboldened to be working with the technology of the future. ⁷¹

Repairing and tinkering: The 1967 incident

The laboratory nature of Chooz also entailed repairing. After only six months of operation, on Christmas Eve of 1967, a control rod jammed after the startup. An initial inspection, including opening a steam generator, found no anomaly, and the plant restarted in early January 1968. By late January, two further rods malfunctioned, revealing a deeper fault. An emergency shutdown followed: inspection of the four steam generators showed severe damage to two, where metal debris had scarred tubes and tube sheets. ⁷² The debris proved to be bolts from the thermal shield inside the pressure vessel (see Figure 6), shaken loose by coolant-system vibrations. The event was classified not as a “nuclear safety” incident – there was no off-site contamination – but as an operational problem.

OPEN IN VIEWER

Figure 6.

The control room of the Chooz A reactor during tests prior to startup in 1966. Most of the plant workers, particularly the French ones, were quite inexperienced.

According to EDF engineers, design flaws by Westinghouse were the cause of the incident. Indeed, very similar incidents, though less serious, had occurred at the Yankee Rowe and Trino nuclear power plants. Wear on the thermal shield was also observed at the Obrigheim nuclear reactor in West Germany. ⁷³ As early as 1968, the working hypothesis was that vibrations in the cooling system were so strong that they affected the thermal shield. In the subsequent exchanges between SENA and Westinghouse, notable disagreements emerged regarding how to address the incident. According to SENA, Westinghouse failed to provide a formal and conclusive demonstration of the causes and sequence of events. SENA considered that the simple replacement of the thermal-shield bolts was insufficient and suggested that the reactor core itself might need to be replaced. ⁷⁴

The repair process, which extended over two years (1968–70), was a haphazard process of constant troubleshooting, tinkering, and trial and error. It involved the delicate dismantling of internal structures and proved to be highly complex due to the radioactivity of the primary circuit affected by the incident. First, the reactor had to be flooded with water so that the staff could work on it without being irradiated. Then, it required developing innovations, including remote-handling tools, such as manipulators and cameras (see Figure 7). Other tools, used outside the nuclear field, were also employed: binoculars, an astronomical telescope, endoscopes, and fibroscopes. ⁷⁵ Pierre Condou, instrumentation technician at EDF, recalls that cameras were mounted on a system allowing vertical movement to inspect each fuel element along its full height and on all four sides. When pieces of debris or metal were spotted, operators used mechanical hand-controlled pincers to remove them remotely. ⁷⁶

OPEN IN VIEWER

Figure 7.

Television camera observation of a metallic fragment on the lateral face of a fuel element. Cameras and television technologies proved to be an unanticipated but crucial part of the repair process.

Lacking any expertise with managing such a serious incident in a large-scale PWR, the French and Belgian utilities relied tremendously on expertise from Westinghouse. Westinghouse specialists, stationed at the Trino nuclear plant in Italy, were summoned to Chooz to help with the repairs. ⁷⁷ However, Pierre Schmitt, who worked as an operator for EDF, felt that the Westinghouse engineers were not very helpful: “[He] treated us like his laborers. Do this, do that, measure this, measure that. And then go on, be a good boy, go on. It was essentially that.” ⁷⁸ Belgians from ACEC, a Westinghouse subsidiary, played a crucial role, too. Tudy Brognon, a Belgian chemical engineer who worked at ACEC, was one of them. Although very new to the nuclear field, his task was to make the water with which the reactor was flooded clearer so that it was easier to do the repairs. He remembers the work as involving a lot of trial and error. After failing multiple times, he and his team finally managed to make the water clearer by using a filter from a sugar factory. ⁷⁹

Donald Berquez was a young engineer and reactor designer when faced with the incident. There were still many uncertainties. He and his colleagues, all very young and inexperienced, had to constantly check and double-check all calculations and information with each other. The engineers from Framatome had to collaborate closely with those from ACEC, a much larger and older company. Berquez often had to travel to Charleroi and remembers it as a “big machine” with its own traditions, and there were some rivalries. A lot of engineering diplomacy was thus required. According to Berquez, relations with Westinghouse were difficult at times: the company was reluctant to share documentation and was not particularly proactive. Indeed, Westinghouse held almost all of the essential know-how for designing PWRs. Many of those who worked on the Chooz A project still remember very well that “Westinghouse was God.” ⁸⁰

Westinghouse and SENA held weekly meetings to address the incident. Despite lacking direct experience with this reactor type, the CEA assumed the role of expert and regulator. ⁸¹ The centrality of Westinghouse to the Chooz project reinforced the CEA’s uncompromising stance toward both SENA and Westinghouse. ⁷⁵ “The CEA was the inquisition,” recalled Donald Berquez, who worked for Framatome and served as an intermediary between the CEA and Westinghouse. He remembered how different the organizational cultures were: the CEA, a hierarchical military–scientific institution, expected meticulous detail, especially for a reactor type it viewed skeptically, while Westinghouse, a large corporation, tended to offer the opposite and, at times, withheld calculations and documentation. Yet the CEA also lacked the technical depth needed to evaluate American designs. The cultural clash was palpable from the first joint meetings, held at the Château de Breteuil. The CEA scientists had hosted the proceedings in a grand setting and wearing formal attire, while the American and Belgian delegates, accustomed to first-name familiarity and a more relaxed business-style approach, felt out of place. According to Berquez, regarding Chooz’s safety, the CEA scientists were also on the back foot during the meetings, contributing little beyond asserting that gas–graphite reactors were inherently superior. ⁸²

Euratom also intervened, creating a tripartite commission of experts from France, Belgium, and Euratom to assess the damage, determine the causes of the shutdown, and bring it to an end. ⁸³ In addition, a consultative committee of European safety specialists, with representatives from Belgium, Germany, Italy, and the Netherlands, reviewed the manufacturers’ repair proposals and even solicited help from the U.S. AEC. ⁸⁴ This European approach, which grossly exceeded the usual remit of Euratom’s Article 37, was unique in the French context. Euratom’s oversight, usually very contentious, seemed more acceptable to the French authorities because of the unprecedented nature of the incident and the specific characteristics of the Chooz A reactor, being an American design not intended to produce plutonium for French bombs. At the same time, the experts were not very transparent about the incident to the outside world. Members of the European Parliament were very much in the dark about how serious the repairs were. ⁸⁵ Still, the moment was significant for Euratom: a year earlier, one of Euratom’s leading officials, Willem Vinck, had unsuccessfully tried to position the organization as a regulator by drafting European nuclear-safety standards. ⁸⁶ At Chooz, Euratom saw another opportunity to play a meaningful role in nuclear safety and assert itself as a regulator.

Decisive meetings in Autumn 1968, nearly a year after the first problems appeared, brought together the CEA, Euratom, designers and manufacturers, and SENA. ⁸⁷ Participants agreed that the incident did not call the PWR line into question and traced the failure not to on-site problems but to defects originating in Westinghouse’s Charleroi works and at Framatome. Westinghouse was then authorized to propose vessel repairs. It recommended permanent removal of the thermal shield. The lack of protection of the steel against irradiation was to be offset by removing fuel elements. Despite reservations, the CEA approved. On September 12, 1968, eight fuel elements and the shield were permanently removed – an unprecedented alteration of a reactor core. ⁸⁸

Ultimately, the reactor restarted on March 18, 1970, after expensive repair works (paid by Westinghouse, ACEC, and Framatome) and an equally costly shutdown of two years. Indeed, the incident had turned Chooz into a laboratory to an even greater extent. For all the actors involved, the shutdown had been an arena for exchange, diplomacy, decision-making, and learning in an unprecedented situation. Westinghouse gained expertise about previously unknown risks across its fleet, and the CEA deepened its understanding of PWRs through inspections and by organizing conferences. ⁸⁹ Belgian utilities and EDF likewise accrued operational experience. ⁹⁰ At Chooz, operators had not been trained for accident management or nuclear safety. Lacking formal procedures, engineers and technicians had improvised solutions on the spot. Euratom, supported by the AEC, sought lessons to prevent similar events in operating or planned reactors. This difficult episode thus left a lasting mark. ⁹¹

Tritium: A laboratory and its problems in the public sphere

Chooz A experienced other unforeseen problems due to its laboratory nature. In the middle of the 1970s, it released larger amounts of tritium than expected into the river Meuse. Tritium is a radioactive isotope of hydrogen and a common by-product in nuclear reactors. All nuclear power plants release some quantity of tritium – albeit limited – into the environment. But the design of Chooz A brought about much larger releases for two reasons. Firstly, reactor designers had added boric acid to the cooling water. ⁹² Boric acid absorbs neutrons. It helps to control the nuclear chain reaction and prevents large fluctuations. Adding more boric acid enabled the operators to experiment with increasing the power output of the plant. But it came with a downside. The reaction between neutrons and boric acid occasionally produces tritium. ⁹³ Yet, there was also a second reason for the high tritium releases. Tritium can permeate the fuel pellets that contain the uranium fuel. At Chooz A, these pellets were made from stainless steel as opposed to the more expensive zirconium alloy that was widely used elsewhere, allowing for much more tritium to leak away. ⁹⁴ The reason for choosing these cheaper materials was to experiment and cut costs.

Although these preconditions caused tritium releases to be high from the outset, controversy only really ensued in the early 1970s when environmentalist movements, which were flourishing by this point, began to scrutinize these releases. This was the first time that experimentation at Chooz faced opposition from public actors. The first sounds of protest came from Belgium. A group of Dutch-speaking environmental activists published a book in 1972 called Zwart Milieu (Black Environment), which quoted research from the University of Leuven arguing that water in the Meuse was contaminated. These results concerned a group of Dutch socialist parliamentarians. Indeed, the Meuse flows through the Netherlands, downstream from Chooz. Meiny Epema-Brugman and Adriaan Oele interrogated the Dutch Minister of Transport and Water about the pollution. Oele, who was also a member of the European Parliament, appealed to the Euratom Treaty to compel the French government to control the pollution. ⁹⁵

This did little to alleviate local concerns along the Meuse, particularly in Belgium. In Visé, a town close to Maastricht, panic ensued in January 1973 after someone saw dead fish in the Meuse. Civil protection authorities immediately proceeded to distribute water bottles, fearing that the drinking water supply may have been polluted. ⁹⁶ The French local newspaper L'Ardennais covered the story for multiple days, much to the frustration of the French prefect of the Ardennes département, who called the newspaper “sensational.” ⁹⁷ But in the Belgian part of the Ardennes, there were similar concerns. La Namuroise, for instance, complained how the region “produced all the water of the country but only the capital [Brussels] and Flanders profit from it.” Together with the construction of new dams in the region, the pressures on the water supply were becoming too high. ⁹⁸ These water debates in Belgium were hardly unique. Across Western Europe, the arrival of nuclear power plants magnified existing struggles over water supply and pollution – up to the point that it became one of the first policy domains in which the European Communities began to engage with environmental policy. ⁹⁹

The tritium levels were determined by the operational capacity of the plant, but also by the water level of the Meuse, river works further downstream, and droughts. ¹⁰⁰ These external factors caused the situation to deteriorate in 1974, as one of the fuel pellets of the reactor ruptured. This caused the cooling water to become heavily contaminated by tritium. ¹⁰¹ Excessive releases of tritium continued throughout 1974. In September 1974, however, Pierre Pellerin, the leading figure of the French radiation protection agency SCPRI (Service central de protection contre les rayonnements ionisants), became concerned with the situation too, particularly because the water level of the already shallow Meuse River was very low. Pellerin kept in close contact with one of the directors of the plant, Jean Grangetas. ¹⁰² Pellerin was known for being very pragmatic. ¹⁰³ He initially granted Grangetas the permission to release multiple times tritium-contaminated coolant water into the Meuse, exceeding the regulations. ¹⁰⁴

However, by the end of the month, no more water could be released. The SCPRI, led by Pellerin, had granted a temporary exception to the normal release of tritium, but even that limit had been majorly surpassed. ¹⁰⁵ Tritium-contaminated water had filled up all the storage tanks, and there was nowhere to go with it. SENA became desperate and warned that the plant would not be able to function properly anymore. Jean Kieffer from EDF complained to Pellerin that the Meuse did not contain a lot of tritium at all and that Chooz A was the only industrial installation releasing it into the Meuse. Kieffer proposed a number of radical solutions, including shipping the water away by truck, boiling and vaporizing it, or even changing the material of the fuel pellets. ¹⁰⁶

The Belgian contingent was surprisingly inactive in this discussion, and there is significant evidence that this was because they were deliberately denied information. At various points in time, the SCPRI was hesitant to provide the Belgian authorities with details about the tritium leaks. Even though Belgian communities downstream would be hit much harder by this contamination than French ones, Belgian authorities were not immediately informed following the release of large amounts of tritium due to the ruptured fuel pellet. In Brussels, the IHE (Institut d'hygiène et d’épidémiologie) studied and supervised radioactive contamination. An SCPRI official noted that it would be “delicate . . . to keep the IHE informed.” SENA was not willing to keep Belgian authorities in the loop either, as they argued that tritium releases were not regulated by the French–Belgian deal that allowed Chooz A in the first place. ¹⁰⁷ The collaboration, diplomatic exchange, and proximity with Belgium thus did not automatically entail a good knowledge exchange. Even more so, such cross-border production of ignorance about nuclear risks was certainly not unique and also took place, for instance, on the Danish–Swedish border. ¹⁰⁸

By 1976, the tritium releases had decreased again. ¹⁰⁹ However, SENA’s experimentations had pushed the boundaries of the regulatory regime governing radiation releases in water. Indeed, the use of cheaper fuel pellets entailed significant river contamination, as well as scientific and geopolitical consequences in the 1970s, aggravated by the border situation and official secrecy.

MOX: A Belgian testing ground

A final experiment that contributed to Chooz’s status as a laboratory was to add plutonium to the reactor fuel. In the 1960s, several laboratories and research reactors in the United States had begun experimenting with MOX fuel. This fuel is a mixture of uranium and plutonium oxides. When an LWR produces energy by nuclear fission, new radioactive isotopes are created as by-products. One of them is plutonium. By reprocessing this plutonium and using it in MOX fuel, some of it can be recycled, meaning less uranium needs to be used. By the late 1960s and 70s, as fears over the scarcity of uranium supplies were growing, scientists and engineers were increasingly taking an interest in MOX fuel. This was particularly the case in Belgium, which, after Congo gained independence, did not have any Belgian uranium or plutonium production. ¹¹⁰

Belgian scientists had been among the first to experiment with MOX fuels. Researchers of the STK (later SCK-CEN) began this as part of the research collaborations with the United States following the Euratom Treaty. To acquire expertise, Belgian nuclear companies funded internships at laboratories in the United States immediately after the Treaty was in effect. ¹¹¹ The very first MOX fuel pellet was built at the STK and loaded into the BR-3 reactor in 1963, in collaboration with the company Belgonucléaire and the United Kingdom Atomic Energy Agency. ¹¹² A few years later, Euratom built a new factory next to the STK, dubbed Eurochemic, where plutonium was produced. ¹¹³ By 1970, Belgian scientists had become eager to test out the MOX fuel produced in Mol in the Chooz A reactor. ¹¹⁴ Through SENA, the Belgian utilities involved EDF, and by extension, the French nuclear industry, in using MOX. ¹¹⁵ They could count on the enthusiastic support of Euratom, and particularly its Belgian top nuclear safety official, Willem Vinck.

However, the MOX experiments had not initially been taken into account when designing Chooz A. A safety review, therefore, had to be carried out for this experiment. Of particular concern was the potential radioactive contamination of the environment, with larger amounts of plutonium isotopes, in case an accident breached the containment. ¹¹⁶ Just as during the incident of 1968, the CEA, together with Euratom, acted as the nuclear regulator. However, there was a major issue. Both institutions had almost no in-house scientific expertise on fuel reprocessing and the related safety risks. Most expertise was centered in the group of companies involved in reprocessing at the STK and Eurochemic, spearheaded by Belgonucléaire. Gianpiero Santarossa, an expert working for the Commission of the European Communities, was concerned about how Belgonucléaire and the other organizations involved would be able to guarantee the quality of the fuel reprocessing. In essence, the concern here was regulatory independence. No one was able to completely verify whether the new MOX fuel was safe, not least the government agencies. ¹¹⁷ In addition, the CEA, as the main nuclear fuel producer in France, was also part of the group of organizations delivering the MOX fuel to Chooz. Again, the MOX experiment exposed the lack of independent regulatory control. This enabled the Euratom Commission to take a leading role in investigating the safety of the MOX fuel. Yet, complete independence remained difficult, as collaboration with the fuel producers, especially the CEA, remained essential for acquiring knowledge regarding the risks of plutonium handling. ¹¹⁸

Just as with the other safety issues at Chooz, the regulatory officials were very dependent on expertise from the United States. In 1974, Vinck traveled to the United States and brought back new information and publications about MOX experiments in the Idaho National Laboratory. ¹¹⁹ That same year, SENA inserted four test elements into the reactor core of Chooz A. ¹²⁰ Two more followed in 1979, and eight more in 1987. ¹²¹ In 1988, SENA sought the approval of the French nuclear authorities, which by now had more extensive expertise in nuclear reprocessing, to recycle the plutonium from spent uranium fuel. ¹²² While the MOX experiment at Chooz, as opposed to some of the other experiments, was a success, MOX fuels never really took off globally. Ironically, despite having started as an American and Belgian project, France is today the only country in the world that still produces MOX fuels. Indeed, the experience at Chooz has proven crucial for fuel reprocessing science and technology in France.

Conclusion

This paper traced back the history of perhaps one of the world’s most curious pieces of nuclear infrastructure: the Chooz A reactor on the French–Belgian border. This unusual nuclear power plant was never simply built to generate electricity or “demonstrate” a certain technology. It also functioned as a laboratory, although not one revolving around scientific experiments but rather technical know-how. It was a place where technologies, expertise, and international collaborations were tested, tinkered with, and sometimes contested. From its cavern design and early experiments for countering irradiation embrittlement, to the crisis of expertise exposed by the thermal-shield incident, to the controversies around tritium pollution of the river Meuse and the pioneering use of MOX fuel, Chooz became a site of experimentation.

This laboratory was transnational in scope and is a prime example of how Kenji Ito and Maria Rentetzi have conceptualized nuclear diplomacy. French, Belgian, American, and European state and nonstate actors all invested in Chooz, seeing it as an opportunity to acquire knowledge that could serve their own national agendas. The Chooz A project demonstrates how transnational collaboration and national interest were never contradictory but deeply entangled in the nuclear industry. EDF had gained experience with designing and operating American reactor technology. Belgian companies had practiced building a reactor on a larger scale while reducing the financial risk. Westinghouse was able to test its reactor design outside of the United States. And for Euratom, Chooz was the first real product of its research and development program in collaboration with the United States, and a litmus test as Euratom stepped up for the first time as a nuclear regulatory agency. Although Euratom is generally described as a failure, knowledge gave the young organization power. Chooz A would very likely not have existed without it. But this transnationalism also had its limits and asymmetries. As historians have shown, transnational collaboration in the nuclear field was fused with national interests and border conflicts. This was clear during the tritium controversy, for example, when Belgian authorities were deliberately kept out of the loop.

As with any laboratory, Chooz’s experiments caused a lot of unintended events and even some problematic consequences. This succession of new problems and challenges made Chooz into a continuous training ground for engineers and other plant workers from France, Belgium, the United States, and elsewhere to learn how to design, operate, repair, and troubleshoot a water-cooled reactor. And how not to. For example, the incident with the thermal shield caused a shutdown of over two years, mostly because there were significant uncertainties as to how to solve the problem. Furthermore, the introduction of MOX fuels into Chooz exposed a lack of regulatory independence. And the particular design of the reactor core caused large quantities of tritium-contaminated water to be released into the Meuse, prompting international indignation. Indeed, this experimentation occurred in a context of mounting concerns about the environment, such as water quality, and growing antinuclear mobilization, which was particularly strong in border regions. Still, these environmental concerns seem to have had a relatively limited impact on the operation of Chooz A in the early years.

Is the case of Chooz A then unique or typical? We would argue it is both. On the one hand, the border location and number of organizations from different countries made Chooz a uniquely transnational laboratory and site for nuclear diplomacy. Other nuclear power plants around this time, like Yankee Rowe or Trino, were also sites of experimentation, but never to the degree of Chooz and in the same transnational context. Chooz should also be seen in a specific temporal context. It is perhaps unlikely that the same degree of experimentation and troubleshooting would have been possible or relatively uncontested a decade later, when antinuclear mobilization and scrutiny were a lot more prominent.

On the other hand, we argue that Chooz A as a laboratory allows us to understand better the role of knowledge-making in shaping worldwide nuclear trajectories, and arguably even energy transitions, more generally. There is a long historiographic tradition of separating nuclear knowledge production from the development of nuclear power generation technologies and facilities. The history of Chooz demonstrates that science, engineering, and practical know-how were much more entangled through the constant experimenting and testing of new technologies and practices within power-generating facilities by nonscientists, such as engineers, technicians, and workers. Power is not always generated according to plan, with solutions sometimes working and sometimes not, like in a laboratory. The picture becomes even more diverse when regulatory agencies and policymakers are brought into the picture, who rely on the same scientists, engineers, and operators involved in the nuclear projects for expertise, but also attempt to independently develop their own knowledge. Chooz, then, is a perhaps extreme but useful reminder – and a pointer for new research – that crucial knowledge production through experimentation does not only take place in scientific laboratories but also in all kinds of energy infrastructures.

Today, France and Belgium are some of the most nuclearized countries in the world. This history of Chooz A is the origin story of that dependence on nuclear power. In the early 1970s, many of Chooz A’s staff had already moved to new nuclear projects, such as Tihange in Belgium and Fessenheim in France, and transferred their knowledge to new groups of experts and applied it on a larger scale. ¹²³ Gradually, the laboratory work at Chooz left a large imprint on postwar science and technology. Yet, in striking contrast to this lasting impact, Chooz A is today forgotten by many, although much more so in Belgium than in France. A decade after its startup, it was already considered by both the Belgian and French nuclear players, who had advanced to undertake larger projects, a technological backwater. ¹²⁴ Today, Chooz A is a laboratory again, albeit as the first water-cooled nuclear reactor to be decommissioned in France, and one of the first in Europe.