Why the international experimental nuclear fusion reactor is in 'big trouble'
theguardian.com
theguardian.com
Let that sink in. Yes, ITER it self isn't ever going to be producing that energy but it is a step in the direction of that end goal.
How can 5 billion dollars extra be considered too much for that end goal? Especially given everything that has happened on the world stage in the last 4 years, and had dramatic effects on the project. Not to mention that the collective yearly budgets of the nations involved is higher than 10 trillion dollars.
Unlimited clean energy.
A personal view on this is also that a rising tide lifts all boats. If the slow moving, very expensive international government project causes more nimble private enterprise to see a way to get there first and get a slice of the budget cake then great!
Unlimited clean energy. It is worth it, IF we can get there.
Multiple sized eggs in different baskets is the best plan.
I bet they even have lots of wind power they could trade.
(2021) https://spectrum.ieee.org/why-the-shipping-industry-is-betti...
(2024) https://fortescue.com/news-and-media/news/2024/03/15/world-s...
4,300 km long extension cord: https://www.suncable.energy/
No, this was never the goal. At the beginning, ITER was a science project to get the US and the USSR to collaborate on something even if the Cold War was at its height. Similar to the International Space Station. Some project with no real scientific objective, but that could spur the enthusiasm of the masses, so politicians could gain popular support while covertly pursuing a path to rapprochement.
> Unlimited clean energy.
How exactly to you get the tritium? By running some fission reactors, right? Fission reactors are a reality today, and can provide unlimited clean energy already. Why don't they? Because they are expensive? And how do we know fusion reactors, if they ever show up, will be inexpensive?
Ok, here's something else to let sink in. The best fusion reaction out there (deuterium+tritium) yields only about 5 times as much energy per kg of fuel as the fission of uranium or plutonium. Only 5 times, not 50 times or 500 times.
And of course, we don't find tritium in nature, so we need to breed it from lithium using a source of neutrons. You could use a fusion reactor as the source of neutrons (deuterium + tritium = helium + neutron), as you can read on this ITER page [1]. What that page does not tell you is that you would need 100% efficiency to be able to breed your own tritium. Good luck with that. Any deficit will need to come from fission reactors.
Somehow the dream that is sold to people is that once we get fusion reactors, we can stop using the "dirty" fission reactors. It is just a lie.
How so? As I understand it, the breeding reaction is exothermic (Li-6 + n -> He-4 + T) so all you need are neutrons, for which (as someone pointed out to me two years ago[1]) neutron multipliers would be used.
As for neutron multipliers, they are currently so far from being realistic that they are essentially science-fiction.
[1] https://www.sciencedirect.com/science/article/abs/pii/S00223...
[2] https://scipub.euro-fusion.org/wp-content/uploads/eurofusion...
Also if all reactors are essentially based on boiling water. Why don't we just dig to depths were water is already boiling? I'm sure there are plenty of those areas on the globe.
Many of the deepest shafts are commercial diamond mines. Not science.
Imagine a Liquid Nitrogen cooled probe through the earths volcanic layers. There are probably organisms living there that we have no idea about.
And isn't that 'four or five times' number, per reaction? Because deuterium and tritium are so much lighter than fissionable materials (by 10X) the energy per kg is truly 50X not 5X. Right?
Per reaction, fission produces between 190 and 200 MeV. Since the input is one nucleus of U-235 or Pu-239 and one neutron, the mass is either 236 or 240, so the yield is 0.8 MeV per Dalton. So the deuterium-tritium reaction is only 4.4 times more energetic than fission, per unit of mass.
As for your first part, I didn't get what you were trying to say.
> Breeder reactors are not some sort of perpetual-motion device.
I don't think I implied they are, and I don't think anyone believes in perpetual-motion devices in 2024.
[1] https://en.wikipedia.org/wiki/Nuclear_fusion#Criteria_and_ca...
Unlimited, clean, predictable energy. If we had spent a few billions on it last year, we'd be using that energy this year already. No breakthroughs needed, everything is already well-known and proven.
Granted, converting hydrogen back to electricity has another loss of about 60% IIRC but still, you get to keep almost half of your electricity when storing it via hydrogen. So as long as solar is at most half as expensive as the second best stable energy source it's a win.
The goal of producing fuel is for explosive, quick release energy. Things like cars, tractors, heavy machinery need fuel. Lighting your home and things like that work fine with predictable trickles of energy that can be produced easily with solar.
The energy density of a fuel like hydrogen means that you need significant surface area to produce usable quantities at scale, quantities that then are burnt quickly. You have to make it at the rate that you use it, which means it's just added complexity in the stack of getting solar energy to where it will be used. It's just solar power with extra steps.
Add to that, solar power isn't this unproblematic perfect solution. It has a whole host of problems. First and most obvious is the mineral requirements to produce it, and extracting those minerals is incredibly destructive. Disposal of waste is also destructive. But another problem nobody talks about that is dismissed because it is seen as a small problem, is the problem of surface area. A limited amount of energy hits the surface of the earth, and life uses every drop of it that it can. Energy production with solar scales two dimensionally. You can't scale this up without significantly damaging ecosystems.
A little box you pump very little resources into that produces absolutely massive amounts of energy, something with a very small footprint that scales extremely well, that's what you want. Solar isn't it. Solar is great, but it's a stopgap for fusion. Everything is a stopgap for fusion.
The only downside is that it’s intermittent, which hydrogen solves. The “extra steps” convert intermittent solar to predictable energy, useable 24/7/365. It’s not a device to transport solar to _where_ it will be used, but _when_ it will be used.
Even if fusion actually worked, it would be inferior in a major way - it concentrates too much energy in a single point. With solar and gas, energy could be distributed and redundant. No sabotage or failure can cut the energy supply to millions of consumers at once, and energy would be created where it’s needed.
Fusion could still be useful for science experiments requiring enormous amounts of energy but IMHO it’s not a good idea as a grid source.
Reality check: We currently have no clean source for abundant tritium.
The 10 year, $6bn project is now 14 years old and already spent $20bn. Where did they get so far? Well they not apparently are confident that it was a 29 year, $60bn project all along. They simply can't know that and it seems likely that we would get further down the road with little to show for all the effort and expense.
By all means people should continue to research fusion, but that should be done on a scale more reasonable than asking for a blank check to build a full reactor when we don't even know how it needs to work. In the meantime, we could make real strides in reducing our total energy use if we really did care about that metric.
We only know it’s doable in the sense that fusion itself has a large net energy release. We don’t know if it’s even theoretically possible (net energy release wise) in the conditions we can potentially create on earth outside of a nuclear explosion.
Way softer metal than Tungsten.
I guess its about the density?
Beryllium Copper is one of the highest temp coppers but that ain't saying much for something that generates millions of degrees of plasma....
Beryllium Copper is tough and used for non-sparking tools. I don't see it handling millions of degrees.
Tungsten seems like a good switch... Imagine machining it though...
What if it crashes into side of vessel?
Tungsten might be good liner but they lose the beam and.......
Far from unlimited. Almost all fusion power plant concepts are thermal power plants. These directly contribute to global warming, and you can’t scale up by more than a couple of orders of magnitude without causing significant climate change.
In the end you’d need to use panels to radiate the waste heat directly to space, but then you have similar land use limitations as solar panels.
There’s no reason to think we actually need that energy. Advanced deep geothermal is probably much easier than fusion. With that you’ve got geothermal near the poles, solar everywhere else and wind/hydro/wave/tide to supplement based on needs and availability.
Helion fusion power plant concept is not thermal. So we can probably make a lot more electricity without heating up the planet too much. In the end it too will be limited by the heat added to the planet.
I think a lot of people aren't aware why ITER was constructed. A big part was to reduce risk. Risk not as in cost, but risk as in failing to generate fusion. It's a big science project and its success will make all the private companies that much better because the knowledge it will produce. ITER is not in competition with private companies in a sense of building a product. ITER never was and never will be a power plant. It is a research device. The only competition is bragging rights of begin first.
The next step is to run self-powered fusion for long periods. We haven’t done either part before. It needs to be demonstrated, studied, and controlled before can build power reactor.
This happens every time with these big science projects. People think they are extremely costly. In one sense they are, but in terms of your tax dollars? They're well within a rounding error. I mean we got over a dozen individuals that could independently build and maintain 10 CERNs (each of them could build 10 CERNs) if we consider the factors of time and compounding interest. Idk, it just does not seem like that much money to me.
[0] The people writing the stories are probably equally dumb, but they have a responsibility to not be
[1] You set out to do something no one has done before, you really think you'll make an accurate estimate? Especially when we have systems that reward underestimating over accuracy. How could this ever happen when we reward the lowest bidder and there are no consequences for being wrong?
[2] US/China/India/South Korea/Japan/Russia each foot 10% of the bill. So ~$150m/yr by naive estimate https://en.wikipedia.org/wiki/ITER#Funding
It’s a bit sad. The online edition of The Guardian has been a sorry shell of what the newspaper used to be since Viner took the helm and brought with her the practice of the US edition. I guess it’s what happens when you put too many people who genuinely believe in post-structuralist theories together.
Which is to say, it is full of statements about ITER I've heard for decades at this point - namely, the illusive "industry will totally overtake ITER".
It's been said before, and it hasn't happened: someone announces a thing, declares scale up in 5 years, 5 years passes and they don't have anything - but at the time all the people super-keen to have "totally called it" and prove how smart they are take a pre-emptive victory lap celebrating the success of the private industry over government science, or ITER or something...just slightly misremembering that it hasn't happened yet. Then it doesn't happen.
This article is citing the US NIF as a potential competitor to ITER: it's not. In any possible way. Or to put it another way - the NIF took about 30 years to get to ignition. They are nowhere near Q > 1 (it's also basically not the goal of the facility - the facility is a nuclear weapons research lab first and foremost).
I will confidently make a prediction here: in 5 years Helion will be dead. Commonwealth Fusion Systems still won't have anything. Tokamak Energy won't have anything. The link in this article about private investment is to another article which is already 7 years old. The website[1] of Tokamak energy talks about fusion "in the 2030s."[1] General fusion are saying break-even by 2026.[2] Tri-Alpha-Energy are saying early 2030s[3] for "Da Vinci" (which I'm just going to assume is their power producing prototype since it's hard to tell from the website). Commonwealth Fusion Systems say they're building SPARC right now as a net energy machine, but as of 2024 they are still building it[4]. You know, the next thing will just be "net energy" as a quick to-do point.
Now...do I wish any of these company's ill-fortune? Not at all. I hope they succeed wildly and it turns out everything works great. We'll be several billion in research dollars ahead if someone cuts off ITER at the pass with a viable, cheaper reactor...but they haven't done it yet. You don't get to take a victory lap about their success when there aren't net energy fusion reactors out there. Because saying you can do it, and actually doing it are quite different - otherwise we'd have succeeded at fusion in the early 1970s when we were writing all the initial grant proposals, not spending the next decades learning about plasma stability, loss mechanisms and scaling rules. After all, people were very confident about Polywell fusion being the next big thing till...whoops, once you have the right model of what's happening it's pretty clear you can't get it to work.[5]
[1] https://tokamakenergy.com/about-us-fusion-energy-high-temper...
[2] https://generalfusion.com/
[3] https://tae.com/fusion-power/
[4] https://cfs.energy/technology#sparc-fusion-energy-demonstrat...
You can certainly imagine that you simply drop pellets into the chamber continuously, and use a heat absorption medium (because you don't have magnets to keep cool) to smooth out the power curve. But all of that is the sort of development work which ITER is designed to prove out for Tokamaks - i.e. everything up to "we have the boiling water now". You'd have to build the ITER-for-Inertial-Confinement reactor project to make progress on it (and a huge chunk of your problem would just be "how the heck do we not spend all our power in the lasers?").
Fortunately we're likely to keep plugging away at it anyway, because the primary goal of NIF is to simulate nuclear fusion reactions in hydrogen bombs without detonating hydrogen bombs - and thus the secondary research can happen anyway. But it's definitely not in the category of "clearly the superior approach" - and more importantly, it wasn't "quick" to do - they've been plugging away at the problem decades.
Just demonstrating Q>1 won't be useful.
If it's improvements in plasma modeling and material science, won't it be possible do achieve that in smaller scale experiments?
Like it's very hard for me to believe that the only way to improve knowledge of plasma and stuff is to spend $20+B on ITER.
Especially as ITER goals were set in 1980s. Computing capabilities are now many orders of magnitude better and are going to grow further in 2030s. Isn't there a possibility that by 2040 they'd just be able to do high-precision simulation of ITER?
And if you have a choice to spend $20+B on building the thing or $20+B to simulate it, the later is probably much more preferable as then you can simulate millions of alternative designs without actually building them.
Plasma is hard to simulate because it is chaotic and includes fluid mechanics and self-generated magnetic fields. There have been effects that have been discovered that weren’t predicted by theory. Simulation can’t be done at finest enough level. We also don’t know the parameters for ignition.
The toxicity of beryllium is well known... how did this design decision go through? Someone was asleep at the wheel.
As for the change, I'm just an interested amateur, so shouldn't go into details and please take this with a big grain of salt, but I've heard it was a result of proposed scenario(s) that some do not regard as realistic.
Let's face it, changing the project name but still calling it ITER is a bit confusing.
(As a former SciAm subscriber - '70's to '90's - I would not look to current-day SciAm for examples of anything good.)
Because that's its name? You can't choose how other people refer to you.
Similarly, when Kentucky Fried Chicken officially changed its name to KFC, nobody cared, and KFC continued to mean Kentucky Fried Chicken in every context except the imagination of some executives.
Successes in one piece of the puzzle, like 2022's success in fusing deuterium and tritium, are paraded around with claims that we just "solved" fusion. In reality the success was a proof of concept that contained fusion is possible, and that it was done while requiring more energy input then was created.
We're 14 years and $20 billion dollars into a project that was mean to take 10 years and $6bn. What have we learned? Plenty of small things certainly, but apparently they also believe to have learned all the unknowns and can now promise that it will only take 15 more years and another $40bn.
I don't buy it, and no one else should either. They don't know how long it will take or what it will cost. 15 years isn't how long the construction process for a now fully understood reactor will take, its a bunch of fluff time hoping its enough to figure out what they still don't know.
When do we say enough is enough and throw in the towel? Unlimited free, clean energy sounds great and all but it also sounds like snake oil. We would be much better off reducing our need for energy rather than continuing to chase a magic solution that we still don't understand how to build.
A bit more realistically, one could describe ITER's top goals as photo ops for politicians, checking boxes for diplomats, and life-long employment for careerists. Wikipedia's summary has conceptual design work ending in 1990, but tokamak assembly not starting 'till 2020. They're hoping for "first plasma" in ~2034.
That is not the timeline of a project where scientific results or technological success actually matter to the folks in charge.
What is the correct time-line for a project that exists at the boundaries of theoretical physics and mega-engineering?
In the history of humanity's major projects, 50 years is not even a blip. Europe's grand cathedrals took centuries to complete.
If your degrading climate and straining electrical grid need economical & at-scale fusion power NOW, that's a rather different situation.
When stimulated by burning military needs, fission technology went from the the first fission experiments in the US in Jan'39, to the Chicago Pile (first nuclear chain reaction) in Jan'42, to military use of fission bombs in Aug'45, to routine operational deployments of fission-powered military submarines (a very demanding application) the late 1950's.
Fusion power is not a realistic solution to global warming. Even in the most optimistic scenarios, where every single current fusion power project hits its declared milestones successfully, fusion power will not be a major part (say, more than 1%) of global electricity production by 2100. Renewables, fission, and degrowth are the only possible solutions to avoid climate catastrophe. Fusion might be a path to greater reliability and prosperity in the subsequent centuries, if we avoid the worse outcomes of global warming.
> When stimulated by burning military needs, fission technology went from the the first fission experiments in the US in Jan'39, to the Chicago Pile (first nuclear chain reaction) in Jan'42, to military use of fission bombs in Aug'45, to routine operational deployments of fission-powered military submarines (a very demanding application) the late 1950's.
Fusion reactions are incredibly more complex than fission. All you need to make a fission power plant is a large-ish mass of fissile material clumped together - it heats up and boils water. The rest is control to prevent various runoffs or runaway reactions.
In contrast, fusion requires inchomprenaibly large pressures applied to a gas to even get one pair of atoms to fuse. Then, the energy of the fusion event tends to push other atoms away, requiring even more pressure to keep the reaction going. The only conceivable way to achieve this is using extraordinarily powerful magnets, and even the most powerful we know how to make require special shapes to actually achieve the required pressures in a relatively tiny volume. Then of course, they need cooling, and the whole structure needs to be solid enough to hold the mass of the magnets.
So, fusion needs special magnets, special mathematical shapes, and numerous engineering challenges to contain all of these. It's not in any way surprising that it is going to take much more time to develop than fission took.
Not to mention, the atom bomb was developed using enormous resources. The paultry 100B dollars over 50 years that ITER will probably take is nothing compared to the Manhattan Project (compared to the GDP of the time). The Manhattan Project at one time employed 133,000 people. Give ITER 65,000 employees and see if they can accelerate their time-line.
Nitpick: Yes, the Manhattan Project got far more money and people than ITER. But:
- ITER is merely one of >100 fusion power experiments - https://en.wikipedia.org/wiki/List_of_fusion_experiments
- Cost/time trade-offs are not linear. The MP's war-priority timeline created enormous inefficiencies. Vs. fusion power research has been going for ~7 decades now. If you add up the budgets of the >100 experiments, over 70 years - then what's the comparison?
You have finite resources
-and-
Excessive CO2 is pushing the only place where your species can live closer and closer to "no longer inhabitable" status
...then how many decades and $billions should you pour into a line of research which shows no signs of real-world utility, after 6+ decades and umpteem $billions of research?
The UK's HS2 project is projected to cost in the ballpark of £50 billion. This is literally a railway line—something that people have been building for a couple of centuries now. Sure, there's a bunch of land acquisition and construction, but it's still a bit of a joke how expensive it's got.
The F-35 project has an estimated lifetime cost of US$1.5 trillion. This is a full two orders of magnitude more than ITER. Sure, this is meant to be spread out over the next 50 years or so, but the initial production cost over-runs were so big it isn't even funny.
Plenty of mass rapid transit systems easily run into the tens of billions of American/Australian/Singapore dollars/pounds sterling/euro.
In contrast to all of these, the money given to ITER is a complete pittance for the sort of science and results that it is expected to generate from the late 2020s. It is a big construction project, similar to a large airport or skyscraper. (Let's not even talk about how expensive Berlin Brandenburg airport got—it's a bunch of bitumen tarmac and a handful of buildings that managed to punch past 6 billion euro.) Its construction has steadily and consistently progressed in the past eight years that I've been paying close attention to it. I am fully confident that it will meet its timeline of first plasma in the late 2020s. Probably closer to 2025, actually.
And the F-35 was the first mass produced Generation 5 fighter and it still is the only one. Definitely not useless for deterring enemies of freedom such as Iran, Russia and China.
Fusion has an alternative: Fission.
Sure, it might not be as sustainable in the long term (although with wide deployment of Breeder Reactors it might be), but we have enough Uranium for the short term anyway.
And we have many decades of experience with them, how to deploy them most economically, how to use them with the least downtime (more than 90% uptime in many reactors) and so on.
As other commenters have said, ITER's budget is pretty small compared to both other similar projects, and to the magnitude of the potential pay-off.