On 9 February 2022 researchers at the Joint European Torus (JET) fusion reactor, based at the Culham Centre for Fusion Energy (CCFE) in the United Kingdom, announced that ‘JET produced a total of 59 Megajoules of heat energy from fusion over a five second period’, during which JET ‘averaged a fusion power (i.e., energy per second) of around 11 Megawatts (Megajoules per second)’.
What inferences were made, based on this result? Producing 59 megajoules of heat energy from fusion over a period of five seconds indicated ‘powerplant potential’ (CCFE), was ‘a significant step forward en route to power generation’ (ABC) and promised a ‘near-limitless source of clean energy’ (Nature).
I will suggest these inferences are hype—not because I presume to know future technological prospects better than fusion scientists but because the inferences violate reasonable standards of good science communication.
The JET experiment (and ITER)
The idea of a fusion reactor is that it will replicate the workings of a star, where hydrogen nuclei collide and fuse into a heavier helium atom and release vast amounts of energy. Minus the gravitational pressure of a star, Earth-bound fusion must utilise a reaction between two hydrogen isotopes (deuterium (D) and tritium (T)) to produce an (electrically charged) helium nucleus (alpha particle) and one neutron. At the extreme temperatures of the envisioned reactor (150,000,000 degrees Celsius; ten times greater than the core of the sun), electrons separate from nuclei and a gas becomes plasma. Sufficient plasma density to ensure collisions occur must be achieved, and sufficient (magnetic) confinement must be deployed to hold the plasma within a defined volume.
The announcement of the JET experiment was greeted enthusiastically. For instance, the CCFE website, Nature, the ABC, Al Jazeera, the Financial Times, Popular Mechanics, New Scientist and countless blogs all directed our attention to the JET experiment being a scaled-down version (one-tenth the volume) of the International Thermonuclear Experimental Reactor (ITER; ‘Iter’ means ‘The Way’ in Latin).
The significance of the JET experiment cannot be detached from the promise of the ITER project. Or, as the director-general of ITER noted, the JET result—‘a sustained pulse of deuterium-tritium fusion at this power level’—is ‘nearly industrial scale’. ITER is the next step in the promised fusion industry.
There are certainly technical similarities between JET and the future ITER: deuterium-tritium fuel, tungsten in the divertor and beryllium (not graphite) in the plasma-facing wall, and remote-handling robotic capabilities.
ITER is a 35-nation collaboration to build a magnetic fusion tokamak reactor, currently under construction in Cadarache, France. Unlike conventional power plants (fossil or renewable) that convert mechanical power into electrical power, a tokamak would draw upon the energy produced by the fusion of atoms to heat the walls of the reactor vessel, which produce steam and then electricity by way of turbines and generators. ITER will produce ‘500 MW of fusion power from 50 MW of input heating power’; that is, it will produce power at a Q value of 10.
The Q value
According to ITER’s head of communication, writing in 2017, ‘to understand the Q of ITER is to understand its most essential operating parameter as well as the raison d’être of the ITER Project’. The Q value is the out-versus-in power-amplification ratio. Specifically, ‘the ratio of the amount of thermal power produced by hydrogen fusion compared to the amount of thermal power injected to superheat the plasma and initiate the reaction’.
ITER is designed to produce plasmas of Q ≥ 10, which is how you get the 500MW from 50MW quoted above. The break-even point, Q=1 (equivalent power out to power in), would to most citizens be a minimum goal for (electrical) power plants.
The Q value of the JET experiment was 0.33.
The JET experiment failed to break even, consuming significantly more power than it produced. As one critic calculated, the reactor lost 98.3 per cent of the energy it consumed—an improvement on JET’s 1997 result of losing 99.4 per cent of the power it consumed—and overall JET’s gross output of 59 megajoules of energy is equivalent to 23 car batteries. When the director-general of ITER claims JET was ‘nearly industrial scale’, the adverb ‘nearly’ modifies the reader’s sense of how close a successful power plant might be. The prospects of success are exaggerated beyond what Q=0.33 might suggest by itself by shifting what counts as sufficient evidence into the future.
Science communication is not a one-way downloading of facts to target audiences but a two-way act of sense-making between audiences. Science communication has multiple goals, the most simple of which is making specialist knowledge claims accessible and generally interesting. Audiences, from general publics to investors to policy makers, also have an interest in science communication that provides both accurate information (reliability) and relevant information (to aid inductive predictions about likely consequences and outcomes of research). More broadly, science communication has the goal of trustworthy information (featuring accountability, integrity and transparency to enable judgements about special interests, uncertainties, risks and benefits).
Much of the media coverage of the JET experiment satisfies the first two norms of producing enticing and interesting stories. Fusion power is accessibly explained, and the meaning of success is translated as a possible energy revolution. Note that guesses about likelihood can be innocent extrapolations as part of meaningful stories without being hype. Indeed, no science communication can be free of exaggeration because it is communicating about an object that is infused by inductive and often politicised practices.
First, scientific reasoning itself is an inductive business that, because of inevitable uncertainties, goes beyond current evidence to generalise about what is likely. Second, because scientific knowledge is not absolute, accepting false things and denying true things will happen, effectiveness will be frustratingly contingent, and qualified and approximate will be the norm, rather than absolute certainty. Third, judging whether sufficient evidence exists for a claim is subject to the inductive risk of being wrong. Assessments of the consequences of error hinge on ethical and political value judgements, ideally in an indirect role of managing uncertainties in evidence and not as direct reasons per se. Fourth, modern political-economic orders demand that techno-science demonstrate relevance to socio-economic questions and contribute to political disputes. In that context science becomes coupled to the media because of the role the media play in framing public opinion, which science taps into in efforts to shape public legitimacy.
To assess if hype is at play in science communication therefore involves more than detailing entertaining and optimistic extensions beyond current evidence. Instead, we should make value judgements about whether information is accurate and reliable enough to assist audiences in making assessments of likelihood, and whether the communication facilitates trust that uncertainties, interests and cost-benefit considerations are being conveyed openly.
By that criterion, we see hype in fusion coverage via a specific mechanism. The future is mobilised as the basis on which to assess whether sufficient evidence is being presented for successful prospects of fusion research.
For instance, at The Guardian a physicist—the author of a book on nuclear fusion—noted that ‘in keeping with Jet’s design and objectives, less fusion power was generated than was needed to heat the fuels…[nevertheless the JET experiment was still a]…compelling indication that bigger and better star machines’ are around the corner. Popular Science opined that ‘at JET, 59 megajoules is .33Q, which is still a step in the right direction’. New Scientist did not report Q=0.33 but probably guessed that the technically literate might wonder, and so wrote ‘JET is getting close to breakeven’. Nature, having breathlessly raised our expectations about fusion power in the first two-thirds of the story, mentions Q=0.33 towards the end. Significantly, expectations about what JET was supposed to achieve are now drastically lowered. A researcher at CCFE is quoted as saying JET was ‘never expected to breakeven’, but if ITER applied the same ‘conditions and physics’ of JET, ITER would achieve Q=10.
Hype loves the future
Note that my claim here is certainly not that exaggeration suggests fraud. Exaggeration can obviously signal fraud, as in the Theranos blood-testing episode and the Hwang Woo Suk cloning case. And ironically, in the originating fusion claim, the infamous ‘Proyecto Huemul’ in Argentina in 1948–51, Ronald Richter convinced General Perón that he could achieve controlled fusion. Richter’s claim of a net positive result was fraudulent.
Instead, my claim is close to what fusion-reactor critics such as Daniel Jassby (2017 and 2018) and Steven B. Krivit (2017, 2022a, 2022b) contend: that fusion reactors might not be what they’re cracked up to be. Specifically, there is ambiguity about fusion ‘power’ where qualifications about plasma power gain are omitted, the degree of net power gain is obscured, power inputs (thermal and electric) are inaccurately reported, the difference between scientific and engineering break evens is elided, fuel supply is overestimated, invested energy and power drains are underestimated, waste problems are ignored, there is unwarranted optimism about supply of coolant water, and (consistent with the general nuclear industry) there is too much innocence about the scaling-up problems with any industrial rollout of a mega-project.
Reports of the JET experiment tend to lower present expectations and raise future expectations, effectively colonising the future. From JET results today to projections about ITER tomorrow, expectations perform their announced futures. What this means is that we tell ourselves there is an easily checkable difference between real and inflated expected value, so that we could check how reliable or true are claims about the future of a technology prior to attempting to build and complete that technology. At the margins of fancifulness, like perpetual engines, the basic ideal of checkability is sensible. But, pragmatically, checking the future of a technology can take us down the same path as trying to build that technology.
That is where fusion power is right now. The imagined future of fusion power is working back onto our present. We can call this ‘hype’ because of a normative judgement that using the future as the basis for assessing present sufficiency of evidence, for future prospects, artfully invites readers to make imaginatively grounded inferences. Doing so falls foul of the spirit of the accuracy, relevance and trust goals of science communication. Of course the hype suits those advancing techno-fixes to climate change, though even there fusion power could be said to only satisfy the cause-effect rule for a good technical-fix and not the rules stipulating techno-fixes should satisfy unambiguous criteria and build on standardized cores. When imagination and materiality start to collide, good science communication should be far more open about the challenges of actualisation, providing a basis for consideration of political-economic power structures and the anthropology of change, so that citizens can make fair projections about techno-scientific futures.
Richard King, 28 Jul 2020
The bright young things of Silicon Valley, with their dreams of direct democracy on Mars and digital immortality, are often difficult to take seriously. But their hubris is only the gaudy version of a broader cultural and political belief in the power of science and technology to edit, alter and override the very stuff from which our world is made—in other words, to ‘play God’.