IN MOST SCIENCES YOU BUILD MACHINES TO ALLOW YOU TO DO experiments. In fusion, the machine is the experiment. You build it to see if it will work and how it works. Because fusion machines take such a long time to build, fusion scientists are always looking one, two or more machines into the future: while they’re building one, they’re always planning more. So it was in the late 1970s, when the construction of JET and TFTR was only just beginning, that many researchers were beginning to think about what was to come next.
The route to fusion energy that had been mapped out earlier in the decade had three stages. First was a demonstration of the scientific feasibility of fusion, in other words a gain greater than 1, and the big tokamaks being built in Princeton, Culham and Naka were expected to take care of that. Next would be technical feasibility, a machine that would produce large amounts of energy while testing some of the technologies that would be needed in a fusion power station, such as superconducting magnets, a system for extracting heat to raise steam, and a method for breeding tritium for fuel. The final stage was commercial feasibility – a prototype power reactor.
So while the three big machines still existed only on paper, some of the more far-sighted planners were already thinking about the even bigger machine, the ‘engineering reactor’ that would come after. One of them was Russian theorist Evgeniy Velikhov. Velikhov was a rising star in the Kurchatov Institute’s fusion department in the 1960s where he teamed up with fellow young theorists Roald Sagdeev and Aleksandr Vedenov – soon dubbed by colleagues as the ‘holy trinity.’ Plasma theory was too narrow a field to contain his talents and he later branched into lasers as well as computers and automation. He was also a skilful political operator – he knew how to play the Soviet system of patronage and political influence. His star was rising so fast that in 1973, at the age of just 38, he took over the reins of Russia’s fusion programme following the death of Lev Artsimovich. In 1974 he was made a full member of the Soviet Academy of Sciences and three years later was elected its vice president.
Before his death, Artsimovich sent Velikhov to represent the Kurchatov Institute in discussions at the International Atomic Energy Agency (IAEA) in Vienna. Ever since the 1958 Geneva conference, fusion scientists had maintained a constant dialogue between East and West. The IAEA organised its regular fusion conferences which all nations could attend. There were visits to each other’s labs and exchanges of information but relations stopped short of any formal cross-border collaboration apart from that of Euratom. That started to change when Velikhov, along with Amasa Bishop of the US Atomic Energy Commission and IAEA chief Sigvard Eklund, formed the International Fusion Research Council (IFRC) in 1971 to advise the agency in its efforts to coordinate worldwide fusion research. Although the IFRC was simply a group of advisers, Velikhov hoped it would prove influential in moving fusion research towards closer collaboration. He already suspected that an engineering reactor might be so large as to be beyond the capabilities of a single country’s fusion research programme.
Velikhov was not the only one thinking along these lines. Within all the fusion programmes, researchers were beginning to realise that as soon as they had cracked the problem of getting a plasma to burn they would have to learn how to handle the neutrons, extract heat and breed tritium. One of those was David Rose, an engineer who was hired by the Massachusetts Institute of Technology in 1958 when it set up its Department of Nuclear Engineering. In the late 1960s he carried out a detailed study of how energy in a fusion reactor would be exchanged between different types of particles – electrons, deuterium and tritium ions, and alpha particles – and how you would inject fuel into the plasma and remove helium exhaust. His calculations suggested that a fusion reactor would be economically viable, but it would need to be big. In 1969 he co-organised a meeting at Culham, the first meeting to consider the engineering issues of a fusion reactor, and this encouraged many more engineers to get involved.
As the 1970s progressed and tokamaks grew bigger and performed better, the need for engineering solutions became more pressing. In 1977 Rose invited senior engineers from different countries to a meeting to discuss how they might better work together. The group weren’t sure how to form an international collaboration but concluded it should probably be organised by the IAEA. It turned out the IAEA was already moving in that direction. The agency’s chief, Eklund, asked the IFRC for suggestions of how the IAEA could take a more active role in fusion research and Velikhov quickly stepped up with a plan he had already worked out: an international collaboration to design a reactor with the express purpose of testing the technology necessary for a commercial reactor.
The project, which was named the International Tokamak Reactor or INTOR, began in 1979. Each of the four participants – Euratom, Japan, the Soviet Union and the United States – nominated four researchers who would meet in Vienna several times a year for workshops that lasted between four and six weeks. When the researchers returned home they would delegate work to their colleagues to be carried out before the next workshop and so the network of researchers involved became very broad.
The majority of fusion researchers didn’t take INTOR too seriously; they were too busy getting TFTR, JET, T-15 and JT-60 up and running. The regular trips that a few of them took to Vienna were just a sideshow, but the INTOR workshops did gradually build up a database of knowledge on how fusion reactors work, taking results from all the fusion programmes. The workshops produced a number of reports describing the theoretical reactor they were working towards, each report increasing in detail and sophistication. But perhaps INTOR’s main achievement was that it showed the very different traditions and methods of the various fusion research programmes could work together and get things done.
But by the mid 1980s it was clear that INTOR wasn’t going anywhere. Although its designs for an engineering reactor were highly praised by researchers, there just wasn’t any political support for an international project to build a giant fusion reactor. Velikhov was frustrated that his ambition to move quickly towards fusion power generation had stalled. The Soviet Union was certainly in no position at that time to build an engineering reactor itself: the economy was in bad shape and work on the T-15 reactor had all but come to a standstill. Velikhov needed a way to get political support behind INTOR and, by an incredible stroke of luck, an opportunity fell in his lap to take his appeal right to the top. That opportunity came in the shape of Mikhail Gorbachev, an old university friend of Velikhov’s, who in March 1985 became general secretary of the Community Party of the Soviet Union – Russia’s de facto leader.
The two had been at Moscow State University at the same time, Velikhov studying physics and Gorbachev law. Gorbachev became active in the Communist Party and on leaving university rose swiftly through its ranks. In the early 1980s, the deaths in fairly rapid succession of Soviet leaders Leonid Brezhnev, Yuri Andropov and Konstantin Chernenko led the Politburo to decide that younger leadership was needed. So, just three hours after Chernenko’s death, the Politburo elected its youngest member – Gorbachev, aged 54 – to the top job. He immediately made his mark as a reformer with his policies of glasnost (openness) and perestroika (restructuring) which sought to loosen the shackles of the old regime. In foreign policy he made decisive moves to reduce East-West tension. He withdrew SS-20 intermediate-range nuclear missiles from central Europe less than a month after taking office and within six months proposed that the Soviet Union and US both cut their nuclear arsenals by half.
As head of the country’s fusion programme and an old friend, Velikhov met with Gorbachev and described to him how an international project to build a fusion reactor along the lines of INTOR, with Soviet and American researchers working side-by-side, could help to diffuse Cold War antagonism. Gorbachev eagerly adopted the idea and in his first foreign trip, to France in October, he discussed the idea with President François Mitterand and received a positive response. Gorbachev’s next foray abroad was to meet US president Ronald Reagan in Geneva at their first summit. In the weeks running up to this November meeting Velikhov worked feverishly with contacts in the White House to get something ready for the leaders to agree on, despite strenuous opposition from the Pentagon which was concerned about valuable software and technology being handed over to the Soviet Union. The often tense and argumentative summit meeting was dominated by discussions of human rights and the Strategic Defense Initiative, Reagan’s proposed nuclear missile shield. No breakthroughs were made in East-West relations but the two leaders did establish a relationship that would stand them in good stead in the future. The rather bland final communiqué did however include – as one of its twelve bullet points – a commitment by the two powers to work together and with others to establish the feasibility of fusion energy ‘for the benefit of all mankind.’
Ronald Reagan and Mikhail Gorbachev at the Geneva summit, November 1985.
(Courtesy of ITER Organisation)
The government agencies responsible for fusion in each country were slow to get moving in forming a collaboration. It was only after the Reykjavik summit between Reagan and Gorbachev in October 1986, when the matter was brought up again, that the bureaucrats got moving. The United States and Soviet Union, together with Euratom and Japan, formed a Quadripartite Initiative Committee to discuss the idea but talks went far from smoothly. Euratom was already well advanced in planning its follow-up to JET, the Next European Torus, and Japan had its own plans too, so neither of these saw the urgent need for another reactor design. US defence officials were still concerned about the transfer of sensitive technology and the Russians wanted any joint design work to be done in a neutral country. After somewhat tortuous negotiations the four partners agreed to work together for a couple of years to produce only a conceptual design – a broad outline that doesn’t get into the detail needed for actual construction. Paulo Fasella, Europe’s representative at the negotiations, gave the project a name: the International Thermonuclear Experimental Reactor, or ITER. Fasella, a highly educated man who had a glittering career in biomedicine before joining the European bureaucracy in Brussels, pointed out that iter in Latin means ‘the way.’
The headquarters for the conceptual design work was Garching in Germany and the project proceeded in a similar vein to INTOR: each partner seconded around ten researchers to the project who would spend several months each year in Garching and then delegate work to their colleagues back home. The collaboration of researchers from four different traditions didn’t always go smoothly. One American researcher described it like this: the Europeans would rant and rave passionately about every issue and the Americans had to explain to the Japanese that they didn’t really mean it; the Japanese would explain their point of view very calmly and quietly and the Americans had to explain to the Europeans that they did really mean it. There were scientific differences too. The Japanese were keen to have a reactor that could demonstrate continuous, or steady-state, operation, while the Europeans wanted the highest possible gain.
In spite of the differences, this project had a different feel to it than INTOR. Because the scientists were now working as directed by their political leaders it somehow felt more real, as if this reactor would actually get built. That dose of realism was also reflected in the design of the machine. When predictions are uncertain, fusion reactor designers tend towards bigger and more powerful machines in the hope that more plasma and stronger fields will swamp any inadequacies. So while INTOR had predicted that a tokamak 12.4m across with a plasma current of 8 MA would be enough to reach ignition, the ITER conceptual design called for 16.3m and 22 MA.
After two years of work the ITER team had come up with a conceptual design that they could all more or less agree on. The question then was, what to do next? The original plan had been to move straight on to drawing up an engineering design – an exact set of plans ready for construction. But the world had changed since ITER had been dreamt up as a way of alleviating Cold War tensions. The Iron Curtain was falling apart and even the Soviet Union itself would soon cease to exist. ITER’s political raison d’être had evaporated and neither was there an economic need for it: energy was not high on the political agenda in the early 1990s. The project’s momentum carried it forward, but it took the four partners two years to decide on how to proceed. A major sticking point was where the design team would be situated. The dying Soviet state and then the new Russian Federation were in such dire economic circumstances that they could make little real contribution to the project apart from scientific brainpower, but none of the other three partners were ready to concede the design team to the other two. They came up with the unwieldy compromise of splitting the team in three: one part in Garching would be responsible for the plasma vessel and everything inside it; another group in Naka would take on everything outside the plasma vessel – including superconducting magnets, power supplies and buildings – and the final group in San Diego would be in charge of overall integration, physics and safety.
By 1992 the engineering design project was ready to go. The teams were given six years to develop the final design. All that was still needed was an overall project leader. There were few people in the world who had the right mix of experience in engineering, plasma physics and the management of large projects. But there was one obvious choice, the person who had for nearly two decades led the design and construction of JET and guided it through the demonstration of H-mode to the first D-T burning in 1991: Paul Henri Rebut. And so Rebut left his beloved JET and moved to San Diego to take the reins of the ITER megaproject.
The designers of ITER were faced with something of a dilemma. This reactor was meant to be a demonstration of technical feasibility but the previous stage in the three-step progress towards fusion power – scientific feasibility – had not quite been achieved. TFTR only reached a gain of 0.3 in 1993 and JET would get to 0.7 in 1997. Only JT-60 managed a gain greater than 1, but that was equivalent gain – what it would have been if tritium had been used. As a result, ITER was saddled with the twin goals of demonstrating scientific feasibility and testing the technologies needed for a power reactor. The problem was that these two goals are not easily fulfilled by a single reactor design.
For an engineering reactor you want a plasma that is stable and quiescent, that will burn for long periods at high gain – mimicking how a working reactor would behave. Such a plasma would be a tool for engineers to test things such as the best lining for the reactor vessel, known as the first wall, so that it will not pollute the plasma and will stand up to years of neutron bombardment. Engineers would also want superconducting magnets because they would reduce the energy demands of a power-producing reactor. They would also want to test blanket modules – sections of the vessel wall – containing lithium which would be converted into tritium fuel by the bombardment of neutrons from the fusion reactions.
But if you haven’t yet achieved gain greater than one, you would ideally want a very different reactor – something that is more like an experimental apparatus than an industrial prototype. You wouldn’t want a reactor with a fixed configuration because you need to try out all possible permutations to get the highest gain. You wouldn’t necessarily want a quiescent plasma; you would want to be able to push it to the edge of stability in pursuit of the best performance. And you certainly would not want complications such as superconducting magnets and tritium breeding blankets that make it harder to interpret results. So, ITER was destined to end up as something of a compromise.
When Rebut arrived in San Diego to take control of the project, he hit the ground running and immediately began remodelling ITER in the image of JET. This is not entirely surprising since JET was the largest tokamak around and had proved very successful. Rebut’s team drew up a design with a D-shaped plasma, similar to JET’s, and with a divertor around the bottom of the vessel, just like the one that at that time was being fitted to JET. The only fundamental difference was the superconducting magnets. Under Rebut’s leadership the already large conceptual design grew even bigger, to a machine nearly 22m across. He increased the number of superconducting magnets to hold this huge volume of plasma in place. The whole magnet system – twenty toroidal magnets and nine poloidal magnets – would weigh a total of 25,000 tonnes, roughly the same as the Statue of Liberty.
Rebut didn’t want ITER to use superconductors. It was possible to achieve high gain with conventional copper magnets but the partners wanted technology that was ‘reactor-relevant.’ Rebut argued that superconductors just made everything more complicated. Superconducting magnets are much harder to make than copper ones and they must be enclosed in a secure container called a cryostat so that they can be surrounded in liquid helium to keep them chilled to close to absolute zero (around -270°C). Because ITER will be producing large quantities of heat from its fusion reactions, the magnets must be shielded or they will warm up and stop operating. So the inside surface of the vacuum vessel would be lined with replaceable steel panels cooled by water flowing inside them.
ITER’s divertor was another key piece of technology because it was the only solid object in direct contact with the plasma during normal operation. Its roles, extracting the exhaust gas from fusion (helium) and absorbing heat, required a material with a high melting point that is able to withstand prolonged particle bombardment. ITER researchers made use of many tokamaks with divertors around the world in their search for the right material. Rebut considered some quite radical solutions to ITER’s many engineering challenges, such as using highly reactive and flammable liquid lithium as a coolant for some reactor components, on top of its role as material to breed tritium for fuel.
The first six months in charge was a gruelling time for Rebut. Every weekend he would fly one-third of the way around the world to the next worksite, spend a week there, then move on again. Despite this itinerant lifestyle he found it difficult to keep up the communication between the three sites. It was a very different operation from JET where he had his whole team around him. Though a brilliant engineer, Rebut was not a natural manager; he found it hard to delegate and so took on much of the design work himself. The international partners didn’t like his style of leadership. They wanted their own researchers, in Naka and in Garching, to play a greater role in the design. ITER was too big a project for it to become a Rebut one-man-show and in San Diego he didn’t have Palumbo and Wüster to shield him from the politicians. The ITER partners wanted a more collaborative approach so after just two years in charge Rebut was eased out.
It was no easier finding a suitably qualified leader than it had been two years earlier, but the person the partners decided upon was not just another European but also another Frenchman. Robert Aymar had led the building of France’s Tore Supra, the follow-on to its Tokamak de Fontenay aux Roses (TFR). But when they asked Aymar if he would lead the ITER project, he said no.
A contemporary of Rebut, Aymar had spent most of his career working at France’s Commissariat d’Énergie Atomique (CEA) on plasma physics. He was inspired by attending the 1958 Geneva conference to pursue fusion because of its possible value to society. By the late 1970s he was in charge of France’s fusion programme and when researchers had done as much as they could with TFR he set out to build Tore Supra. This reactor would be the first large tokamak to use superconducting magnets. With the strong steady magnetic fields they produced, Tore Supra would be able to hold its superheated plasma for minutes at a time instead of the seconds possible in a conventional tokamak.
Aymar realised that for such an ambitious project to be successful he needed to have all of the team together in one place, not scattered around universities and labs across France. Just as was happening at JET across the channel, he needed a team dedicated to this one goal and nothing else. During 1984 and ’85 he persuaded some 300 families to move down to Provence as Tore Supra took shape at a CEA lab complex in Cadarache. Aymar and his team completed Tore Supra in 1988 and in the process created a powerhouse of fusion research at Cadarache. Recognising that success, the CEA offered him the job of heading its whole basic physics division. Aymar was in his element: with a staff of 3,000 scientists he was now responsible for the commission’s work in many areas beyond plasma physics, including nuclear and particle physics. But then ITER came calling.
At that time, in 1994, the pinnacle of fusion achievement was TFTR’s D-T shots with a gain of around 0.3 – hardly a convincing demonstration of fusion. In a few years’ time JET would get closer to break-even but to Aymar’s eyes fusion research had a way to go before it could confidently demonstrate high gain. That being the case, the machine that Rebut had been designing in San Diego was the wrong sort of machine. It was too much like an engineering reactor and didn’t have the flexibility that may prove necessary to achieve the physics goal of high gain.
Despite his earlier refusal, ITER’s backers contacted Aymar again and asked him to reconsider. Aymar thought hard about it. Although he was enjoying himself running the wide array of fundamental physics at the CEA and was concerned about the current ITER design, perhaps it was his mission in life to guide fusion a bit further along the road to real power generation. He accepted the job, but he wasn’t happy about it. As soon as he was on board he set off to visit the three ITER work sites – Naka, Garching and San Diego. His aim was to steady nerves and build confidence among the researchers after the change in leadership.
But holding together an unwieldy design project would soon prove to be the least of Aymar’s problems as, in autumn that year, the Republican Party took control of both houses of the US Congress. Two months after the Congressional elections, in January 1995, in a conference centre just outside a wintry Washington, DC, Anne Davies, then head of the Department of Energy’s fusion programme, told the directors of America’s fusion laboratories to begin preparing for the worst.
At that time, DoE was spending $350 million per year on magnetic fusion, but once construction began on ITER and Princeton’s proposed Tokamak Physics Experiment (TPX) much more money would be needed. Building ITER alone would consume more than the whole of the current budget. America’s fusion leaders didn’t have to wait long for the axe to fall. Later in the year, when Congress set the 1996 budget, it awarded magnetic fusion just $244 million. This was barely enough to keep America’s domestic fusion programme going, let alone pay for an expensive new reactor to be built somewhere overseas.
Anne Davies and her DoE staff now had the unenviable task of deciding which programmes would survive and which would not. Lab directors and university department heads started jockeying for position to ensure that their reactor or research project was not cut. ITER, as the most expensive item on the wish list, started to attract the wrong sort of attention. Some in the US fusion community agreed with Aymar and others that the leap to an engineering reactor was too risky and believed something more modest should be tried first. With limited resources, some argued, why should the US be supporting a hugely expensive machine in another country – whose success wasn’t guaranteed – while denying fusion scientists at home money to carry out any meaningful research? ITER also had few friends in Congress. It was typical for a large project to be championed by the senator or member of Congress representing the state or district where the project will be built. ITER didn’t yet have a site, and it almost certainly wouldn’t be in the United States, so it lacked such a champion.
But funding and politics weren’t ITER’s only problems. In the mid 1990s two researchers from the Institute for Fusion Studies at the University of Texas in Austin, William Dorland and Michael Kotschenreuther, developed a new computer simulation of the plasma inside a tokamak and it produced some very unwelcome news for ITER. The very hottest place in the plasma – where fusion is most likely to happen – is the centre, with the surrounding plasma acting as a layer of insulation slowing down escaping heat. That’s one of the reasons why a bigger tokamak is better than a small one, because there is more insulating plasma around the hot core. Turbulence is the enemy of this insulating effect because it mixes plasma from the hot core with cooler outer layers, helping heat to escape towards the edge. Tokamak designers knew about this turbulence effect and used scaling laws to extrapolate from known amounts of turbulence in existing reactors to predict how much turbulence there would be in future reactors such as ITER.
In contrast Dorland and Kotschenreuther predicted in a detailed way how plasma would behave in tokamaks of different sizes and under different conditions. They verified their simulation by adjusting it to mimic existing tokamaks and it was able to predict how they behaved pretty well. When they presented their simulation and its predictive power at conferences, other researchers were impressed. Then the pair applied their model to the proposed design for ITER and received a shock. The simulation predicted that in the large volume of plasma in ITER there would be a lot of turbulence, more than was predicted by scaling laws. This turbulence would bleed heat away from the plasma core to such an extent that, Dorland and Kotschenreuther estimated, ITER might not be able to achieve temperatures necessary for fusion.
When they presented these new results the reception was far from warm. ITER was a multi-billion dollar project to which many researchers had devoted their whole working lives; they were not going to see it sunk by a pair of computer geeks. The simulation was now subjected to much more intense scrutiny and its authors to harsh criticism. The Texas pair stuck to their guns and to this day fusion researchers remain divided over whether their predictions are correct – a working ITER will be the final proof. The research did not derail ITER but it did provide valuable ammunition to the project’s critics.
* * *
By 1997 the ITER team was putting the finishing touches to the reactor’s final design, described in a huge 1,500-page report. The machine it described remained largely along the lines laid out by Rebut: the plasma vessel was 22m across – more than two and a half times the size of JET – with enormous, and costly, superconducting magnets to hold the plasma in place. The price tag of $10 billion was equally awe-inspiring. The target was to steadily produce 1.5 gigawatts of heat – one hundred times JET’s record-breaking output in 1997. And while JET required 14 MW of external heating to keep the reaction running, the ITER design called for 150 MW of heat, from both neutral particle beams and radio-frequency heating. The plan was for ITER to also achieve ignition – heating itself with the energy from alpha particles created by fusion reactions. At that time, no reactor had even got close to ignition.
As the scale and cost of the proposed ITER started to become apparent, US participation in the project was more and more difficult. Ever since the election of the Republican Congress in 1994 the Department of Energy’s fusion budget had been repeatedly chipped away, leading to the cancellation of TPX and the closure of TFTR. Republican representative Jim Sensenbrenner, a fusion sceptic, was appointed chair of the House Committee on Science and Technology, which was ultimately responsible for US fusion funding. Fearing the demise of fusion research in the US, DoE officials were applying tremendous pressure on Aymar and his team to keep the cost of ITER down. Aymar himself was beginning to have doubts about the design, that it was too big, too expensive and too great a leap from current knowledge. On the quiet, Aymar asked researchers at the Garching site to start work on a more modest design, one that would still generate a lot of power and so show that fusion energy was feasible, but also one that was an evolution from existing tokamaks rather than a revolution.
Others were thinking along similar lines. With only $50 million a year to spend on ITER, Davies at the DoE knew that the proposed machine was far beyond America’s means. DoE officials suggested scaling back the ITER design to build a smaller reactor, with less ambitious goals and a price tag cut in half – a plan dubbed ‘ITER Lite.’ Some US researchers, aware of the fact that funding ITER would probably squeeze fusion research spending inside the US down to nothing, were proposing an even more radical retreat: abandoning the idea of a single huge machine altogether and instead upgrading existing tokamaks and carrying out an international campaign of research to better understand burning plasmas. With doubts about turbulent heat transport in the back of their minds, they argued that if there is that much uncertainty it’s probably best not to build a huge, expensive reactor that could still fail. At academic conferences, among US researchers at least, the talk in the corridors suggested that ITER was all but dead.
According to the schedule, now that the design was complete, the partners should choose a site for the reactor in 1998 and then start building, aiming to complete ITER in 2008. But it soon became apparent that the US was not the only partner having money troubles. Japan had long been one of ITER’s most enthusiastic participants and many expected the machine to be built there. But in the spring of 1997, Japan no longer had the money to match its enthusiasm. At that time Japan’s postwar economic miracle had ground to a halt. Often referred to now as Japan’s ‘lost decade,’ the 1990s saw the country try to spend its way out of recession with ambitious public works. When this failed to get the economy moving, the government was forced to tackle the budget deficit by cutting spending. Along with other big-ticket science projects, Japan was forced to cut back on ITER, and so it asked for a three-year delay in the start of ITER’s construction.
For Jim Sensenbrenner and the House science and technology committee this just reinforced their view that the ITER project was in terminal decline. According to the committee ITER, at $10 billion, was too expensive; it was questionable whether it would work (c.f. Dorland and Kotschenreuther); and it couldn’t even be considered a viable project since it didn’t have a site. It also made the committee uncomfortable to be giving so much taxpayers’ money to a project that was not controlled by the US. The committee allowed US membership of the engineering design collaboration to run its course, but from 21st July, 1998 US participation in the project was stopped. Scientists working in Garching and Naka returned to the US. The worksite in San Diego was closed down and non-American researchers sent home. US researchers were forbidden from participating in ITER activities or meetings, even as observers. The unprecedented East-West and then global cooperation that had existed in fusion research for the four decades since the 1958 Geneva conference, and that had outlasted the Cold War, was terminated by order of the US Congress.
The ITER project was in a critical condition. With Europe now the only one of the three remaining partners that still had a sizable fusion programme it was hard to see a way forward. Some in Europe argued that they should just abandon ITER and make something more affordable, a European successor to JET. In Japan, there was a crisis of confidence. Ever since the Second World War, Japan had taken a lead from the US in matters of foreign policy. With America out of the picture, could they trust the Europeans?
The partners set up a working group to consider two options: a large ITER-like machine able to study both the science and engineering of a burning plasma, or a number of smaller machines that could address different issues. The group concluded that the only way to examine how all the many questions surrounding burning plasma interrelate was in a single integrated machine with long pulses and alpha particles as the dominant source of heat. The Japanese decided to stay on board, so the teams in Garching and Naka were set the task of designing a new, smaller machine, costing half as much as the 1998 design but retaining as many of ITER’s technical goals as possible.
Aymar’s hunch that a slimmed-down ITER might be needed proved prescient and much of the necessary redesign work had already been done at Garching. In 2001 the teams presented a new final design for a reactor with a vessel that was 16.4m across, instead of the original 22m, and capable of carrying a plasma current of 15 MA, down from 21 MA. The output power, at 500 MW, was also one-third the previous goal, but the biggest sacrifice was that the new ITER was no longer expected to achieve ignition. Instead of running on alpha particle heating alone, the reactor was likely to need at least 50 MW of external heating to top up the alphas and keep the plasma burn ticking along. That still meant a gain (Q) of 10 but, yet again, one of the major milestones of fusion energy seemed to be retreating out of reach.
The redesign did reduce the cost of the project to a slightly less eye-watering €5 billion, but now the partners had to come to terms with the reality of choosing a site and building it. While the design of the reactor had been left for the scientists to decide – within the limits of the budget available – the choice of site would be a largely political decision and all the researchers could do was sit tight and hope for the best. The plan was for the nation that was chosen as host to shoulder the greatest share of the construction cost – because of the economic benefits of having the reactor on its territory – with the remainder divided equally among the other partners. But only a small part of each partner’s contribution would be in the form of cash paid over to the yet-to-be-created ITER organisation. Most of it would be contributions ‘in kind’: components for the reactor that would be manufactured by each partner’s home industries and then shipped to the site. ITER managers would carefully divide up and parcel out the construction work so that each partner made an appropriately sized contribution while its industry learned the skills that will be needed to build future commercial fusion power stations. Everyone wanted a share of the knowledge that could turn into a multi-billion dollar industry. But the question remained: who would play host and would it be a welcome boon or a crippling burden?
The surprise first entrant into the site contest in June 2001 was Canada, which was not at the time a member of the ITER project. The offer was being promoted by a consortium of companies led by the power utility Ontario Hydro. It had a site on the north shore of Lake Ontario just east of Toronto, next to the Darlington nuclear power station, which was already licensed for construction of a nuclear plant. In many ways, the offer made sense: Canada has plentiful supplies of tritium fuel because it is a by-product of its home-grown Candu fission reactors; Ontario Hydro would benefit by supplying electricity to the project; and, situated halfway between Europe and Japan, the Darlington site presented a compromise solution that might also lure Canada’s southern neighbour to rejoin the project.
But other partners were not ready for Canada to waltz in and carry off the prize. In Europe, Germany had always been the most enthusiastic supporter of ITER and had considered offering to host it, but the cost of reunification with East Germany after the fall of the Berlin Wall meant that it was now not so keen. France, however, had an almost ready-made site: Cadarache, the Commissariat d’Énergie Atomique lab that Aymar had built up into a fusion powerhouse. It had available land and supplies of power and cooling water already built for Tore Supra, plus support from the French national and regional governments. Japan was weighing up three potential sites: Tomakomai on the northern island of Hokkaido, Rokkasho at the northern tip of the main island Honshu and Naka, north of Tokyo, home to JT-60. Russia’s economic malaise still ruled it out as a potential host.
Back in the United States, fusion researchers and the DoE were busy trying to figure out what to do next. Many wanted to get back into ITER as soon as possible and they were encouraged by the fact that the remaining ITER members were working on a smaller and cheaper design. But for a while at least, rejoining the project was not politically possible. The community began working on a design for a new home-grown reactor. Known as the Fusion Ignition Research Experiment (FIRE), the reactor would study the physics of ignition and was championed by Dale Meade, the tall and affable former deputy director of the Princeton fusion lab. Meanwhile Bruno Coppi of the Massachusetts Institute of Technology put forward another alternative: a reactor called Ignitor which followed the model of MIT’s Alcator tokamaks in using very high magnetic fields to get strong confinement and heating. His Ignitor proposal would go all-out to show that ignition was physically possible, but others considered it would do little else to aid progress towards a power-producing reactor.
In July 2002 a few dozen senior figures from the US fusion community gathered at the Snowmass ski resort in Colorado for a two-week-long conference to consider what they should do next. Ultimately it was up to the DoE and Congress to decide what to back, but the scientists knew that if they presented a united front they were more likely to get what they wanted. There were three options on the table: rejoin ITER, build FIRE or build Ignitor. There were vigorous debates, interspersed with walks in the mountains, but a poll at the end showed where the researchers’ hearts really lay: they voted 43-to-1 in favour of rejoining ITER, with FIRE held in reserve if that proved impossible – even Meade voted for ITER. An advisory panel to the DoE considered the same question later that year and all members of the panel voted for ITER with FIRE as a backup, apart from Bruno Coppi who voted for Ignitor.
There remained the problem of persuading politicians that the project they so decisively rejected a few years earlier was now researchers’ top priority, but that now didn’t seem as impossible as it once would have. The Republican Party was not the dominant force in Congress that it had been and then there were the terrorist attacks on 11th September, 2001 which caused a change in the tenor of American politics. George W. Bush had replaced Bill Clinton in the White House in January of that year and his administration did not seem any more enthusiastic about fusion than Clinton’s. The highlight of his early energy policy seemed to be a proposal to open up the Arctic National Wildlife Refuge to oil drilling companies. In the aftermath of 11th September, everything changed: suddenly energy security was high on the agenda. How could the US ensure its energy supply in the event of another large terrorist attack or conflict in the Middle East? Research into energy technologies that didn’t rely on imports of fossil fuels were suddenly flavour of the month.
Just six months after the attacks, with fusion scientists now unambiguously backing ITER, government officials began to look into how the US could rejoin the project, reportedly at the suggestion of President Bush himself. In his state-of-the-union address in January 2003, the President made a commitment to develop cleaner energy technologies and generate more of it at home rather than importing oil. Two weeks later a US delegation travelled to St. Petersburg for an ITER council meeting to begin the process of rejoining. The US was not alone in its new enthusiasm for ITER: China and South Korea also joined the collaboration in 2003.
In the space of five years the fortunes of the project had made a dramatic turnaround. In 1998 ITER had been on the brink of collapse. Now it had six partners – one of whom encompassed most of Europe – several sites vying to host it and a design that everyone believed in. At this high point, Robert Aymar decided it was time to hand the leadership over to someone else. He had supervised the completion of two designs and navigated ITER through its greatest crisis: now another project in trouble had come calling. At CERN, the European particle physics lab near Geneva, construction of its giant particle-smasher, the Large Hadron Collider, was well over budget and struggling to stay on schedule. So they called Aymar. He said at the time that he was too old to embark on a job like constructing ITER, which would go on for another ten years.
Once the question of the site was settled, the ITER project would be transformed into a fully fledged international organisation charged with building the reactor. That would mean new leadership, so in the meantime Aymar’s deputy, the Japanese plasma physicist Yasuo Shimomura, was made interim director. The whole ITER operation was in a state of suspended animation as the team waited for politicians to decide where it would be built. Japan whittled its roster of proposed sites down to one, Rokkasho, while Europe had acquired a second one, Vandellos near Barcelona in Spain. The ITER council, a biannual meeting of delegates from the partner governments, looked into the merits of the four sites and declared them all suitable from a technical point of view. If all of them would work, how were they to choose? The backers of each site began pushing other attributes, as if selling package holidays, to try to persuade other partners: Rokkasho will have Western-style housing for staff and an international school for their children; Cadarache has the weather and ambiance of Provence and the nearby Cote d’Azur; while Darlington is a stone’s throw from the cosmopolitan city of Toronto. A date was set for the council meeting that would make the decision: December 2003 in Washington, DC.
The EU decided that it had to choose between its two candidate sites – Cadarache and Vandellos – to increase its chance of success in Washington. Technical assessments concluded that either would work. Building in Spain would be cheaper but the site, next to an existing nuclear power station, didn’t have any science institutes nearby. At Cadarache there was already a wealth of scientists and engineers on hand if help was needed, but the site was far from the sea so the transport of large and heavy components could be tricky. Debate over the two sites raged on through the summer and autumn. Favours were called in; backs were scratched; and different countries lined up behind their favourite sites. Although the process was divisive at the time, it helped to cement Europe’s determination to win ITER. In a sense, the Europeans felt that they had earned it. While the US fusion budget was slowly whittled away and the Russian and Japanese programmes were undermined by their struggling economies, Euratom had kept the whole project afloat, especially during the years after America’s withdrawal. Hosting ITER on European soil would be the payback.
The person who found himself in charge of Europe’s effort to win ITER was Achilleas Mitsos, a gruff economist from Greece and a specialist in European integration who joined the European Commission, the EU’s civil service, in 1985. As is the custom at the Commission, every few years he was moved to a different job, managing such issues as social and economic cohesion, education and training, and socio-economic research, before eventually becoming director-general for research in 2000. By now a seasoned operator in the Brussels bureaucracy, Mitsos was unruffled by the tussle between Cadarache and Vandellos. It had required some diplomacy: to appease Spain when Cadarache was chosen, it was promised that the organisation that would eventually be needed to manage Europe’s part of ITER construction would be based there. But that did little to prepare him for what was going to come next.
The Washington meeting of the ITER council was to be the project’s turning point, the moment it changed from an idea on paper into an international collaboration intent on building a fusion reactor. President Bush was on standby to come in and lend authority when it came time to sign an agreement. Everyone expected a deal to be done, but the events of 9/11 cast a dark shadow over the meeting. The Iraq war had begun only nine months earlier and relations between the United States and France were in a deep freeze because of French opposition to the war. Now the EU had the temerity to come to the negotiating table with a plan to site ITER in France.
Tensions began to simmer before the meeting even got started. Shortly beforehand, an unsigned document was circulated to all the delegations apart from Japan describing the merits of Cadarache as well as many claimed shortcomings of Rokkasho, including the high cost of labour and electricity, risk of earthquakes and lack of infrastructure. The Japanese were furious. The US, following its recent return to the collaboration, was determined to get the site decision settled at the Washington meeting and so tried to calm frayed tempers.
The partners had been expected to have a sober debate and then come to an agreed position on where to build ITER. In the event, the meeting was a train wreck. First, the Canadians withdrew their site. Ontario Hydro and its partners had failed to win the support of the Canadian federal government and without that the site, and Canada’s membership in the collaboration, were non-starters. With Cadarache and Rokkasho now head-to-head the partners lined up behind their favourites: Russia and China supported Cadarache; Korea and the US favoured Rokkasho. The negotiations became a slanging match. Europeans accused the US of only supporting Rokkasho because it couldn’t stomach giving ITER to France, while the Americans charged Europe with blackmail after some French delegates said that if ITER went to Japan, France would pull out of the whole project. In the end, nothing was decided. The two sites were asked to provide more technical information to help resolve the issue. The champagne stayed corked and the teams headed home amid an air of mistrust and accusation.
For the next eighteen months, fusion scientists looked on in horror as their cherished project became ammunition for diplomatic warfare. Senior researchers were now excluded from meetings deciding the fate of ITER as government officials took over. After the conference-room sniping of the Washington meeting, salvos for and against the two sites continued in the media. US energy secretary Spencer Abraham told Japanese business leaders: ‘From a technical standpoint you have offered the superior site.’ The French prime minister Jean-Pierre Raffarin fired back: ‘We have to have ITER, even if we do it ourselves… We won’t let go of this.’ High-ranking politicians toured the capitals of other ITER partners trying to win support. Both sides hinted that they would consider going ahead anyway with any partners that wished to join them. Japan and the EU even offered to shoulder larger and larger proportions of the total construction cost, in efforts to buy support.
Each side tried to exploit the weak points of the opposing site. Rokkasho’s Achilles’ heel was earthquake risk. Japan sits on the Pacific ‘ring of fire,’ an area around the ocean margin that is prone to quakes and volcanoes. Although Chinese officials didn’t say anything publicly because they didn’t want to inflame historic tensions between the two countries, they felt the seismic risk of Rokkasho was too great and hinted they would pull out if that site was chosen. Cadarache had the problem of being inland, so to get many of the reactor’s huge components to the site, the project would have to widen roads, strengthen bridges and modify junctions along a winding 106-kilometre route. Japan argued that transporting such components by road was impractical if not impossible, but France cited the example of the Airbus A380 superjumbo. Although assembled in land-locked Toulouse, some of the plane’s enormous parts, including whole wings and fuselage sections, are built elsewhere in Europe, shipped to southwest France and then trundled 240 kilometres through the countryside at night on purpose-built transporters.
Meanwhile, officials in Japan and France gathered more information on the suitability of the two sites in nine subject areas. A meeting was held in Vienna in mid March to debate the results. In France, licensing of the reactor would be covered by existing legislation and was already well underway; licensing in Japan required new legislation, a process which had not yet begun. The risk of a large-scale earthquake damaging to the reactor was considered twenty times more likely in Rokkasho than in Cadarache. Estimates put the cost of preparing the site at Cadarache at around one-eighth that at Rokkasho and the mild Mediterranean climate of Provence was certainly more appealing to researchers than the cold winters of northern Japan. In total, Cadarache was considered the better site in seven of the nine categories and in one they were judged equal. Its one failing was its inland position.
The comparison seemed to clinch it for Cadarache but the supporters of Rokkasho refused to concede and, because of their objections, the comparative document was never made public. The United States continued to insist it was supporting Rokkasho for technical reasons whereas in reality US officials backed Japan because they thought it would make a good, committed host, while it could not say the same of Europe. Some in the Bush administration didn’t believe that the European Union could be treated in the same way as a sovereign state. Its twenty-five members had differing levels of commitments to ITER and the administration didn’t think they could act with the unity of purpose that was needed to manage ITER’s construction and pay for it.
Something had to give in the negotiations. The open hostilities were getting nowhere. Both sides realised that there was no way to resolve the issue while one side came out the ‘winner’ and the other the ‘loser.’ To save face, there had to be some prize that would go to the side that didn’t get ITER. So began a series of bilateral negotiations between the EU and Japan, and the topic under discussion came to be known as the ‘broader approach to fusion.’ Fusion researchers had long recognised that to make progress towards a fusion power plant there were other things they needed to do apart from building ITER. They needed a particle accelerator facility to test the radiation hardness of materials that would be needed for such a plant, and supercomputers to simulate it. At that time, no one had any firm plans to build these facilities but now they were needed as bargaining chips. In order to cool the rhetoric between the two sides, the issue of who would have ITER and who the other facilities was put to one side – negotiators only referred to the ‘host’ and the ‘non-host.’ The hope was that if the facilities included in the broader approach became enticing enough, one side wouldn’t mind having them rather than ITER.
The barrages of rhetoric calmed down as the potential host and non-host talked to each other. Mitsos was travelling to Tokyo twice a month during this period. Soon an appealing deal was worked out: the host would pay for almost 50% of ITER’s total cost (with around 10% each from the other partners) and the non-host would get one or more expensive facilities from the ‘broader approach’ whose cost would be shared by the host and non-host. The problem was that both Japan and the EU still wanted to be the host. Something else was needed to tip the scales towards the non-host so that one of the two would be prepared to accept it.
One day in summer 2004, Rob Goldston, the director of the Princeton Plasma Physics Laboratory, was tidying his house. Japan’s deputy science minister was coming to visit the lab and Goldston had invited him to dinner at his home on the evening before the visit. Once everything looked suitably welcoming, Goldston sat on the stairs and tried to think of ways to break the deadlock over ITER’s site because that evening’s dinner provided him with a rare opportunity. Goldston knew that there is a strict protocol when dealing with Japanese politicians and some topics of conversation are off-limits. But there is also an unwritten rule that late in the evening, after a certain amount of alcohol has been consumed, it is acceptable to speak frankly and broach difficult subjects. Goldston had made a list of his ideas and when he was joined by his son Jake on the stairs he showed him the list. Jake, a university economics student, told his father, with the confidence of youth, that they were all useless, apart from number 4.
Idea number 4 was that as an added incentive for the non-host, the host – which is paying for 50% of the whole machine – would pay for some its components for ITER (for example, 10% of the total) to be built by firms in the non-host country. So the host still pays no more than 50%, but the non-host’s industry gets more of the benefit. This, explained Jake, was the only one of Goldston’s ideas in which the non-host got something unambiguously worth having in addition to the broader approach facilities. Goldston phoned officials at the Department of Energy and explained that he wanted to propose this idea to his Japanese visitor. He was told that it would be OK, so long as the idea was not attributed to the DoE.
The evening went according to plan. As soon as enough wine had been drunk, Goldston broached the subject of the ITER site and explained his idea. At the end of the evening the minister went away with a two-page memo spelling out the plan that Goldston had prepared earlier. When the visitor arrived at the lab the next morning, he immediately asked to use a fax machine. He wanted to send the memo to Tokyo before work ended that day. The ball was set in motion and would soon gather more momentum.
It took many more months for all the details to be worked out, but in May 2005 the Yomiuri Shimbun, a Japanese daily paper, quoted government sources as saying that Japan might be willing to give up its bid to host ITER if it won a lucrative role in construction. It took one final EU-Japan meeting the following week in Geneva to seal the deal. The two sides had resolved to have the issue settled before the 6th July start of the G8 economic summit at Gleneagles in Scotland. So at the beginning of July, just days before world leaders would gather in Scotland to discuss climate change and aid to Africa, and George Bush would collide on his bicycle with a British policeman, delegations from the ITER partners were welcomed to Moscow by Evgeniy Velikhov, the same person who twenty years earlier had persuaded Mikhail Gorbachev to propose ITER as a worldwide project ‘for the benefit of all mankind.’ Now he oversaw another turning-point in its progress: as everyone now expected, Cadarache was announced as the site for the reactor.
Also revealed was how much Europe had to pay to get it. The division of the €5-billion cost followed the plan worked out between the EU and Japan including the extra 10% shifted from host to non-host outlined in Goldston’s idea number 4. In addition, 20% of ITER headquarters staff would be Japanese and the EU would support Japan’s proposal for a director general. As for the broader approach, Japan would get to choose a facility to build on its soil, up to a cost of €800 million, with half paid by the EU.
After eighteen months of often rancorous negotiations, everyone seemed pleased with the outcome. Japanese industry could look forward to lucrative contracts paid for by Europe, while Europe could bask in the prestige of being the home of ITER. Although some European researchers worried about what they had taken on: the huge cost of ITER now threatened to starve all other fusion projects. Nevertheless, there was a palpable feeling of relief for everyone involved.
Just over a year later, France was able to chalk up one more minor victory over the United States when ministers from the seven partners (India joined early in 2006) came together to sign the international agreement that would make ITER an official collaboration. The ceremony was overseen not by President Bush in Washington but by President Jacques Chirac at the Elysée Palace in Paris. Earlier that year, Mitsos had stepped down from his job at the Commission and returned to Greece. His job was done. ITER was no longer a dream: it was a genuine collaboration of nations representing – with the addition of India – more than half the world’s population. It now had a staff, a headquarters, a large patch of bare earth, and a plan. All they had to do now was build it.
A 1:50 scale model of the reactor at ITER headquarters in Cadarache.
(Courtesy of ITER Organisation)