“For more than a century, scientists and engineers have dreamed of bringing to earth the same kind of clean energy produced in our sun: Fusion!.”1 Opening narrative in “Fusion is Now” video produced by the the U.S. ITER group at Oak Ridge National Laboratory
Nuclear power can be said to have been born on December 2, 1942 when a team of scientists and engineers led by Enrico Fermi initiated the world’s first controlled chain reaction in a small reactor at the University of Chicago. Thirteen years later nuclear reactors were powering U.S. submarines, and two years after that the first U.S. nuclear generating station started delivering electricity to the grid. By 1992, just fifty years after nuclear power was born, fission reactors were providing 20% of U.S. electricity. There was never any doubt about whether turning atomic fission into electricity could be accomplished. No fundamental engineering challenges needed to be overcome, and no new materials needed to be developed. Fission reactors were basically just a new way to boil water.2 Other than the isotopic enrichment of uranium and the chemical separation of plutonium, both of which were sufficiently understood such that the initial processes worked as predicted. The physics had been worked out in top-secret pursuit of weapons with unimaginably destructive power; after that, harnessing nuclear fission for electricity was mostly a matter of straightforward engineering.3 This is not to diminish the many technical problems encountered and the innovative solutions developed in designing and building fission reactors, but rather to point out that there was never any doubt that a working power reactor could be built.
There were, of course, some problems. All commercial reactors in the U.S. are scaled-up versions of reactors designed to power submarines: compact reactors with high power density. But light-water reactors are not inherently safe (the reactor can melt if not constantly and carefully managed). Although nuclear engineers knew how to design inherently safe reactors in the ’50s, the higher-risk naval reactors had a strong head start and were therefore faster to commercially exploit, so that’s what utility companies chose to build.4 Passively safe reactors that do not need human intervention to prevent meltdown have a much lower energy density— larger size per unit of power— and are therefore not suitable for naval power systems. Looking back, quickly locking into naval reactor designs for civilian power was a short-term expediency that we should now regret. The civilian nuclear industry was not as tightly managed as the nuclear navy, and a couple of high-profile accidents, along with a concerted anti-nuclear campaign by fossil fuel companies and consequent environmentalist opposition,5 For specifics, see Atomic Insights https://atomicinsights.com/smoking-gun-aec-told-president-kennedy-coal-industry-opposed-nuclear-energy/ led to widespread fear of nuclear power— and dashed any realistic prospects of nuclear reactors replacing coal-fired power plants.
Hopes for nuclear fusion, on the other hand, seem to be even stronger now than in 1954 when Lewis Strauss— then chairman of the Atomic Energy Commission, and in apparent reference to fusion power— infamously declared that nuclear power would one day be too cheap to meter. Indeed, nuclear fusion is widely regarded as the best, and some believe the only, hope for achieving widespread, carbon-free electrical power. But optimism about fusion energy appears to come more from desperation than from realistic assessment. Although renewable energy such as wind and solar photovoltaic could theoretically provide all of our electricity, in practice that will require an impracticable combination of production over-capacity, transmission line redundancy, and massive-scale batteries or other energy storage systems— all to ensure a constant flow of electricity to the grid.6 A stable electrical grid requires that backup power be immediately available for intermittent energy sources (such as wind) that contribute more than a small percentage of power to the grid. The large increase in renewable energy production over the last couple of decades was largely enabled by the availability of natural gas generators and other dispatchable sources of energy.7 Dispatchable power plants such as natural gas generators and hydroelectric can be quickly turned on and off in response to grid demand. As the demand for electricity inexorably rises, however, we will need a steady supply of reliable power that is unlikely to be met by renewables.
With coal-fired power stubbornly holding its share of total world electrical generating capacity,8 The Guardian, April, 2022 and the planet rapidly getting hotter, it’s easy to see why many people hope that fusion will allow us to continue our electric-intensive lifestyles without the guilt of contributing to global warming. Visions of a fusion future have been promoted by media reports that almost invariably paint a glossy picture of fusion as the future source of energy— clean, safe and virtually unlimited fuel. In a segment typical of reports on fusion power, for example, Science Magazine recently stated that “Fusion holds the tantalizing promise of plentiful, carbon-free energy, without many of the radioactive headaches of fission-driven nuclear power.“9 Science Magazine, online, Dec. 13, 2022 https://www.science.org/content/article/historic-explosion-long-sought-fusion-breakthrough
Fusion advocates often contrast “safe and clean” fusion with “dangerous and dirty” fission, emphasizing that fusion reactors cannot meltdown and will not produce extremely long-lived radioactive waste. But that is a misleading comparison. It is true that unlike today’s light water reactors, fusion reactors cannot meltdown and do not create long-lived radioactive waste. But many if not most thoughtful advocates of nuclear power believe that the current fleet of fission power reactors should be replaced by more robust, fail-safe designs and a fuel cycle that reuses spent fuel— thereby eliminating the problems of reactor failure and eons-long storage of radioactive waste. Fusion advocates also typically avoid the fact that fusion power is very much a nuclear technology, and that fusion reactors will generate large volumes of radioactive waste that must be isolated, very likely on-site, for up to 100 years.
For these and other reasons the typical narrative presented to the public about fusion— a reliable source of clean energy with virtually unlimited fuel— is at best misleading and at worst outright false. It is simply not true that fusion power plants will not generate radioactive waste or that the fuel will be virtually unlimited; it is misleading to say that they will be safe; and it is premature to conclude that they will be reliable— not to mention that fusion power plants will be extremely, and possibly prohibitively, expensive. Even prominent scientists get it wrong: Michio Kaku, a physicist popular for explaining physics on television, said that “hydrogen from seawater could be the basic fuel. So this is too good to be true.”10 In an interview with CNBC following the announcement in December 2022 that the National Inertial Fusion center had achieved “breakeven” power. He was right about it being too good to be true, at least in part because he was wrong that the fuel will be readily available.
Althought a shortage of fuel is the most damning indictment against the prospects for fusion power anytime soon (more on this later), a large array of fundamental engineering problems must be solved before the lack of fuel becomes the primary issue. We’ve known since the idea of fusion power was first proposed some fifty years ago that new materials, technologies, and extremely complex systems would need to be developed and tested before a functional fusion reactor11 That is to say a reactor that can maintain power for long enough to qualify as a reliable source of electricity. could be built. But the full nature and extent of the complexity and critical compromises that will be required have only become apparent as the technology has moved from academic projections to actual engineering— even if it’s only in a very early phase.
The plasma in a fusion reactor, ten times hotter than the core of the sun, must be kept away from the reactor wall by magnets that are powerful enough to lift an aircraft carrier and supercooled to almost absolute zero— and that are just a meter or so away from the plasma. The intense heat and radiation from the plasma will damage the reactor wall in ways that can only be modeled now, as there is no means of testing them in the conditions they’ll actually encounter. An array of new, technically complex, and interdependent technologies must be developed to harness the energy of high-energy neutrons that create fusion energy,12 80% of the energy in tritium-based fusion comes from high-energy neutrons (the same energy as neutrons released in a neutron bomb). to recycle and breed tritium for reactor fuel, and to maintain continuous plasma. Difficult compromises must be made between materials in critical systems and components, not only for system effectiveness but also for human health and environmental protection. And if a fusion power plant ever becomes operational, maintenance and repairs will have to be done by remote control, including components that weigh many tons.
These are but a few of the problems that are supposed to be addressed by the world’s largest fusion reactor, currently being built in southern France. The International Thermonuclear Experimental Reactor (ITER) is a multi-national effort originally intended to demonstrate the viability of fusion as worldwide source of power. That ambitious objective has been greatly scaled back, however; the current and more modest objectives of the ITER are to test the “availability and integration of the technologies essential for a fusion reactor;” to investigate and demonstrate sustained plasma (although only for less than an hour); and to demonstrate that a fusion reactor can be operated safely. As challenging as these goals are, they are a relatively low bar in the vision of widespread fusion power.
But even if the ITER is successful in achieving these goals, the prospects for fusion power will still be very unclear, as the objectives of the ITER are modest compared to what needs to be done if fusion reactors are ever to be more than the world’s most costly experiment. The task of demonstrating fusion as a viable source of electricity is presently planned for ITER’s successor, the Demonstration Reactor (DEMO). Still in concept infancy, with many fundamental design issues hinging on results from the ITER, DEMO is intended to be the world’s first fully operational fusion reactor, a robust power plant that will demonstrate the technical and economic viability of fusion power. And that is where the big problems with fusion are likely to emerge.
The DEMO power plant, if it is ever completed, will be the most complex, expensive, and temperamental power system ever built. At this very early stage of idea conception, DEMO is expected to produce 800 MW of electrical power, about 80% of what an average nuclear plant produces today. The reactor is expected to be substantially larger than ITER, with huge consequent increases in cost.13 The current expectation of fusion engineers is that reactor costs will increase disproportionately to increases in power output. Tritium recovery in DEMO will require a processing facility about 30 times bigger than the largest tritium recovery facility in the world today. Moreover, it will be a far more complex operation, having to recover tritium from a much wider range of materials and system components, and to do it with a recovery efficiency that is ten times greater than what is achieved in the current state-of-art facility.
Although we don’t have any idea how much all of this will eventually cost, it seems much more likely than not that even if fusion plants achieve sustainable power, they will be far and away the costliest source of energy, both to build and to operate. Some experts believe that the operating costs alone for a fusion power plant will be so cost-prohibitive that the electricity it produced would be far too expensive even if the plant didn’t cost anything to build.
Daunting as these challenges and problems are, the biggest elephant in the fusion room is that the fuel is extremely scarce. A largely-ignored fact of fusion power schemes, and a major misrepresentation by many fusion advocates who should or do know better, is that unlike the sun, fusion reactors cannot burn elemental hydrogen.14 The sun “burns” hydrogen through the tremendous pressures produced by its immense gravity. Because it is effectively impossible to achieve those pressures on earth, any fusion reaction here must be produced with one or more of hydrogen’s isotopes. Fuel for fusion reactors is therefore made from one or both of hydrogen’s isotopes: deuterium (2H) and tritium (3H).15 A hydrogen atom has one neutron; deuterium two, and tritium three. Deuterium can be extracted from seawater, so it is readily available anywhere— hence the often repeated but misleading statement that the fuel for fusion power is unlimited. But deuterium-deuterium (D-D) fusion requires temperatures and pressures that appear to be unattainable with present or foreseeable technology. Deuterium-tritium (D-T) fusion, on the other hand, occurs at temperatures and pressures that, although hotter than the core of the sun, has been shown to be achievable with today’s technology, even if only for very short periods of time.16 D-D fusion requires temperatures of 400-500 million °C; D-T fusion occurs at around 150 million °C. The current record for sustained plasma is about one minute. A 50/50 mixture of deuterium and tritium (D-T) is therefore the only fuel option for major fusion power programs.17 Several fusion projects are pursuing to D-D and other non-tritium reactions, but they are all at a very early stage of development and do not appear to be have much prospect for commercial appcation in the foreseeable future.
But unlike deuterium, tritium is extremely rare: the worldwide inventory of tritium is tiny (about 25,000 grams in 2022) and decreasing by about 5% each year (tritium has a half-life of 12.3 years). It is possible that the first full-scale fusion reactor (DEMO, by current plans), will not have enough fuel to start up, let alone run for more than a few months.
Created exclusively in nuclear reactors, almost all of the world’s accumulated tritium has been produced to make thermonuclear weapons (hydrogen bombs).18 Trace quantities of tritium are also produced naturally in the upper atmosphere. The only non-military tritium presently available for potential fusion fuel is a waste product from CANDU reactors in Canada.19 CANDU reactors use heavy water (deuterium) as a moderator; when a deuterium atom absorbs a neutron it transmutes into tritium. Tritium reduces reactor efficiency, so some countries have built facities to remove tritium from the heavy water. South Korea also has a non-mitary tritium extraction facity, but Korea is shutting down it’s reactors and no longer has commercial quantities of tritium. Canada’s Darlington tritium extraction facility (the only such facility in Canada) produces around 1,500 grams of tritium per year.
The lack of tritium is a huge problem for the world’s fusion program at large. ITER alone is estimated to need about 1,000 grams of tritium per year, most of Darlington’s annual production. But the ITER won’t start using tritium until after 2036, by which time the current world inventory will have shrunk by 50 percent and output at Darlington will have significantly decreased as reactors are taken offline. Far worse for fusion prospects, the DEMO— ITER’s planned successor and currently the most likely first operational fusion power reactor to be built— is expected to use about 300 grams of tritium a day. And that’s just for a single reactor; dozens of other programs worldwide are also expecting sufficient D-T fuel when (or if) their designs achieve sustained operation.
Even if there is enough tritium to start all the D-T cycle reactors hoped or planned to begin operation in the next few decades, fusion reactor designers and operators will still face two related challenges stemming from the extremely limited quantitites of tritium: 1) recovering virtually all unreacted tritium from the reactor “exhaust”, and 2) “breeding” as much tritium as the reactor burns or otherwise loses. During each “pulse” of plasma in a tokamak reactor only about 2% of the tritium will actually be fused, the rest is exhausted from the reactor and must be recovered and recycled into new fuel. Although recovering hydrogen from a fusion reactor’s exhaust stream is relatively straightforward (compared to fusion’s many other far more challenging proccesses), it will be complicated by the fact that every component in the recovery system will become radioactive through the diffusion of tritium into most system components. Recovering tritium from plasma exhaust could also be greatly complicated by the presence of beryllium, a highly-toxic metal that is the material of choice for the plasma-facing layer of the reactor.20 Berylum has been planned as the plasma-facing material for the ITER, but it may be replaced with tungsten due to health and environmental concerns– a change that could have significant consequences for reactor performance and first wall integrity.
Beryllium was selected as the plasma-facing material for ITER because it has a high neutron-multiplication factor,21 The neutron-multiplication factor is an expression of the ratio of number of neutrons released to the number absorbed. an essential requirement for breeding tritium– which is in turn mission-critical for any hope of producing power from D-T fusion reactors. If fusion power plants cannot breed and recover at least as much tritium as they burn, the fusion power program will quickly run out of fuel (assuming there is enough to even start the reactors). Tritium breeding schemes, however, are conceptual at this point and will be tested only minimally in the ITER.22 Tritium breeding has a very narrow window for success and multiple interdependencies between key systems and materials, only a few of which will be tested at the ITER.
Breeding and recovering sufficient quantities of tritium will be a major industrial effort in its own right. Not only must a sufficient quantity of tritium be produced through processes that have not yet been tested at anything approaching scale, it must then be recovered and recycled into new fuel— again using processes that have not yet been developed. Assuming that those processes are successfully deployed, the scale of tritium processing is enormous. A single 1,000 MW(e) fusion power plant will have to recover as much tritium in a 3-5 days as the largest tritium extraction facility operating today recovers in a year23 The Darngton Tritium Removal Facity in Canada, the largest such facity in their world, produces 1-2 kg of tritium per year.– and do so with recovery efficiencies that are ten times greater.24 Environmental laws and regulations usually specify the total amount of a radioactive material that can be released from a facity, regardless of throughput. Larger throughput will therefore require a smaller percentage of material that is allowed to escape to the environment.
Each D-T fusion event releases a single neutron, so in addition to creating heat each neutron must “breed” an atom of tritium in order to theoretically break even on fuel (that is, not to need tritium from outside sources in order to maintain reactor operations). In practice, however, it is not possible to achieve 100% recovery of tritium. Fusion reactor designers must therefore come up with ways to increase the number of neutrons available for creating more tritium.
Lithium enrichment is another new industrial process that must be developed if fusion power is to become more than a hugely expensive experiment. All tritium breeding schemes require massive amounts25 The DEMO reactor is presently estimated to require around 20 tons of enriched 6Li for breeding tritium. of highly-enriched 6Li , a stable isotope of lithium produced by nuclear countries for thermonuclear weapons. But 6Li has not been produced in large quantities for over 50 years. The U.S. stopped producing enriched 6Li in 1963 because the lithium enrichment facility had “lost” about 300 tons of mercury in a decade or so of production (during which it produce a reported 442 tons of enriched lithium26 “Mercury Releases from Y-12 Lithium Enrichment”. The COLEX lithium enrichment process used at the Y-12 plant continues to be used by China and Russia. No other means of enriching lithium has been tested at the scale required for fusion reactors (see World Nuclear Association, “Lithium”).). China and Russia are the only countries currently enriching lithium, but in quantities far too small for even a single fusion reactor. Moreover, environmentally-acceptable alternatives for enriching lithium are currently only concepts, and none of them have been demonstrated to be feasible on an industrial scale. Without 6Li , there will be no commercial fusion power, but the issue is currently either glossed over or ignored by major fusion programs.27 ITER ignores the low worldwide inventory of 6Li and offers a very misleading analysis of expected 6Li requirements for commercial fusion plants. And a recent report by the Institutes of Engineering, Science, and Medicine presenting a strategic plan for a U.S. fusion power plant by 2050 does not mention 6Li production as an issue that needs to be addressed. And there is very little room for error: the entire scheme for fusion power as currently planned requires fusion reactors to breed more tritium than they use, as losses during production and processing are inevitable.
And then there is the matter of fusion as an economically viable source of power. Today’s nuclear power plants, almost all of which have been running for decades using technology developed in the 1950’s, average full-power generation (referred to as capacity factor) of more than 90%— industrial wind turbines run at about 30%.28 Capacity factor is the total amount of power generated by a power plant in a period of time divided by the power that could have been generated in that period if the plant ran continuously at full power output. Nuclear fission power plants achieved rapid commercial success because the technology was relatively simple and straightforward. Many incremental improvements have been made over the seven or so decades of nuclear power, but the basic schemes for commercial nuclear plants has remained the same. Stodgy as they may be, North American nuclear power plants continue to crank out power more reliably than any other source of electricity.29 https://www.statista.com/statistics/183680/us-average-capacity-factors-by-selected-energy-source-since-1998/
It doubtful, however, whether fusion power plants will be able to achieve anywhere near the same level of reliability. Far from being the relatively simple technology of nuclear fission, fusion power plants will be highly complex systems requiring an array of interdependent, first-of-kind technologies— many of which cannot even be tested until they are completed and deployed in a full-scale power plant. But even if all of the fundamental materials and engineering problems can be solved; even if plant operators are able to maintain sufficiently reliable power output; and even if enough tritium somehow is available to operate the reactors, fusion power plants will still face the daunting challenge of economic viability. Fusion power plants are almost certain to be far more expensive to build than today’s nuclear plants and much more expensive to operate. Although a thorough assessment of probable operating costs has not been performed (and cannot be until most or all of the fundamental materials, processes, and systems are settled) it is difficult to imagine that fusion power plants can achieve anything close to economic viability.30 That is, delivering reliable power at a price that is competitive with other low-carbon energy sources. Utilities are no longer buying the relatively inexpensive (by probable fusion plant standards) conventional fission reactors. It seems extremely unlikely that they will buy into a much more expensive and far less reliable technology— as tokamak fusion reactors almost certainly will be.
It is a matter of conjecture as to why the world’s major fusion programs are proceeding without apparent alarm about the shortage of tritium, or why so few commentators include such warnings in their otherwise rosy assessments of fusion energy. But the facts are as they are: a fusion energy program that uses D-T fuel will not produce any meaningful amount of electricity unless it comes up with a lot more tritium. We need to face the likelihood that fusion reactors will not be a viable source of power in any timeframe that could help stave off catastrophic global warming.
This is not to say, however, that the prospects for fusion energy should be dismissed— only that we should take a longer view, carefully evaluate the options, and invest our resources in technologies that could actually pay off, even if not in this century. In such a view, advanced, fail-safe fission reactors appear to be the primary, if not the only, credible option for a medium-term31 Medium term being more than a decade but less than many decades. alternative to fossil fuel power plants. Although a program to build advanced fission reactors could take decades, we know that they will work as planned. Looking several decades down the road, closing the fission fuel cycle (such as the Integral Fast Reactor Program) could in theory provide 100% of our electricity for hundreds of years using the waste from today’s reactors as fuel– and greatly reduce the need for mining (and, more importantly, for enriching) uranium. Having wasted several decades fretting about the impossible goal of risk-free nuclear power, it’s time to confront the fact that all sources of power have risks and environmental consequences and begin to responsibly balance the risks and rewards of all energy options.
For fusion research, the potential for aneutronic fusion must be seriously considered. Aneutronic fusion— where very little of the energy comes from neutrons— has been explored since the middle of the 20th century. If aneutronic fusion could be accomplished (a questionable outcome, to be sure), the problems associated with today’s fusion schemes— high-energy neutrons, tritium shortage, material degradation, and radioactive waste— they all simply disappear. It is, however, much more difficult to reach energy break-even with aneutronic fusion, as the required temperatures and pressures are far beyond what we can achieve with today’s technology. But that is what research is all about: pushing beyond the limits of what we are currently able to do. The illusions of the currently fashionable D-T fusion program should not diminish our imagination of what we may one day actually be able to do.