New design could finally help to bring fusion power closer to reality | Hacker News Reader

New design could finally help to bring fusion power closer to reality

phys.org

127 pointscurtis49 comments

Meanwhile, MIT's fusion program is likely to be shut down soon. Their Alcator C-Mod has the strongest magnetic field of any tokamak in the world, and made a serious breakthrough in tokamak physics several years ago.

I had a chance to visit a couple years ago. A grad student showed us a metal tie, about a meter long, and said they'd calculated that two of them could hold down the Space Shuttle during launch. To hold the reactor together when they switched it on, they needed 38 of them.

They hadn't run the reactor for a year due to funding issues.

I had to run the numbers on that Space Shuttle thing, just to see. The Space Shuttle weighed about 4.5 million pounds at launch, and had a thrust-to-weight ratio of 1.5, meaning there was about 2.25 million pounds of net upward force on it. So each of those metal ties was capable of carrying a bit over a million pounds of load. Neat.

And in Si units about: 500 000 kg (453 592.37)

kg is a unit of mass.The Si unit of force is the Newton.

What's the SI unit of pedantry?

Easier to understand than the American/Imperial one :)

The big problem with controlled fusion, as I understand it, isn't creating a net energy positive fusion reactor like this article implies. The physics is already there, we just need the engineering to catch-up.

The challenge is the materials that compose the reactor become brittle and highly radioactive far faster than usable. Controlled fusion as an energy source isn't just challenging because fusion is difficult, it's also challenging because materials don't handle neutron bombardment well.

I think where this advancement is going to really help is in the iteration phase due to lower build costs but I'm still waiting for a solution to the radioactive economic issues associated with the materials these reactors would be constructed from.

It seems like they address this issue as well:

"Another key advantage is that most of the solid blanket materials used to surround the fusion chamber in such reactors are replaced by a liquid material that can easily be circulated and replaced, eliminating the need for costly replacement procedures as the materials degrade over time. "It's an extremely harsh environment for [solid] materials," Whyte says, so replacing those materials with a liquid could be a major advantage."

It's been half a decade since I studied anything fusion related but IIRC regardless of using a liquid blanket you still have to have the machinery (powerful microwaves, particle injectors, etc) and the inner wall present.

Liquid breeder blankets are definitely the way to go but I don't think they are the full solution. (They are super neat because you can use them to make more fuel for fusion through neutron capture, so not only do you reduce radioactivity but you also reuse the energy in a useful fashion)

> "The challenge is the materials that compose the reactor become brittle and highly radioactive far faster than usable. Controlled fusion as an energy source isn't just challenging because fusion is difficult, it's also challenging because materials don't handle neutron bombardment well."

Isn't that problem solved with aneutronic fusion?

https://en.wikipedia.org/wiki/Aneutronic_fusion

Of course we should focus on getting fusion working (working in the sense of providing more energy than we put in), regardless of whether we use the aneutronic approach or not, but as the field develops I'm optimistic we'll find ways to have robust fusion reactors.

I can honestly say I'd never heard of that before you posted your link. I don't know enough to form a valid opinion here.

"I don't know enough to form a valid opinion here.".

Beautiful. This is pretty much the last thing I ever expected to read on the interweb.

Thank you for your honesty.

I'm not an expert in this field either, but I have been following the development of the Focus Fusion reactor being developed by Lawrenceville Plasma Physics, which is where I was introduced to the concept of aneutronic fusion. This video is a fairly decent introduction to their approach (note that the plan is to use the standard deuterium fuel first as it's cheaper and easier to work with, but the design can potentially support an aneutronic fuel such as pB11 in the future):

https://www.youtube.com/watch?v=O4w_dzSvVaM

There are also reactions with fewer overall neutrons such as D-D fusion, which one of my professors was a large proponent of.

The problem is, low neutron fusion reaction cycles require a very high pressure/temperature/plasma density to reach peak reactivity, substantially higher than D+T and D+D fusion.

If we do eventually progress to aneutronic fusion test reactors, we will first have to get neutron-emitting fusion reactions working to the level of substantial net power generation, even if a given test reactor only lasts a few months before neutron irradiation becomes unmanageable.

According to their paper [0], the entire vaccum vessel is considered disposable and might survive only a few months. (It's also explicitly an experimental reactor meant to study neutron radiation damage, so they're hardly claiming it's a practical commercial design. But I think you're right, that the problem is basically unsolved).

    This allows the vacuum vessel to be replaced quickly,
    mitigating first wall survivability concerns, and permits a
    single device to test many vacuum vessel designs and
    divertor materials.
    
    [...] The replaceable vacuum vessel is made of corrosion-
    resistant Inconel 718, which maintains high strength and
    corrosion resistance at elevated temperatures.
    
    [...] Little research has been done regarding how Inconel
    718 responds to the irradiation environment of a fusion
    device [82]. However, studying the response of components to
    fusion neutron effects is part of the motivation for
    ARC. [...] It is unknown if Inconel 718 would behave
    similarly in a fusion neutron spectrum, but one expects the
    vacuum vessel would survive for at least 6-12 months (15-30
    DPA).

[0] http://arxiv.org/abs/1409.3540

Thanks for pointing that out. I don't think I conveyed myself well in my original post, this is a fantastic development but I feel like the article didn't emphasize that the development is really about moving towards the material science that we need. It felt like the article sensationalized the energy output per unit size and didn't focus enough on the material science challenges of the problem (as well as how much this will aide in their solution).

The authors clearly recognize this, I just wish the public understood it better.

The article was written by MIT's public-relations office, so interpret that as you like.

One thing mentioned in the article was another advance in this design, where some of the materials that deal with the most intense neutron bombardment are replaced by a liquid that can be pumped out and circulated. Presumably there's still a containment vessel for this liquid that is a maintenance pain, but they do seem to be working on the problem.

The liquid blanket technology is nothing new or unusual and, from what I recall, seems to be generally agreed as the way to go. You are exactly correct about the containment vessel, and there is also the raw machinery needed to maintain the containment field and the reaction. I think the lifespan for these things is on an upper bound of 3 years, and I can't remember if that is at power plant or experimentation scale (experimental reactors run way less frequently and create less radiation than a power plant would).

It's an interesting problem to me because fusion is largely solved and well understood from the physics perspective (which is what most people consider the challenging esoteric part), the remaining challenge is a much more 'mundane' task of developing economic methods to create the needed materials.

> "It's an interesting problem to me because fusion is largely solved and well understood from the physics perspective (which is what most people consider the challenging esoteric part), the remaining challenge is a much more 'mundane' task of developing economic methods to create the needed materials."

On one level, yes, it's a well understood problem. I mean when you get to the point when Make can teach you how to build your own fusion device I think it's safe to say we have a fairly decent handle on what's involved:

http://makezine.com/projects/make-36-boards/nuclear-fusor/

However, the challenge of producing a nuclear fusion device that generates more energy than it consumes is still a technical challenge. Perhaps it'll just be a question of better materials, but from a layman's perspective the plasma confinement/control aspect does not appear to be a completely solved problem.

This strikes me as the general problem with nuclear power: it works so well -- in a sense -- that it tears apart the fabric of matter itself, making its components, fuels, wastes, etc. absolute hell to handle. It's tough to use matter to contain matter-decomposition machines.

Not a nuclear physicist, so this is my armchair layman's "why don't they just..." analysis.

But doesn't nuclear fission work off of neutron bombardment? Why not build a hybrid reactor where the "shield" undergoes fission while the core is generating fusion and perhaps they can produce a useful amount of power, perhaps with the fission "shielding" powering the magnets or some such?

What would be the advantage of a hybrid like that over a straight-up fission reactor?

checks the box for fusion?

more seriously, it could split the energy production to some ratio of fusion/fission and produce less hard to handle waste. Since the fusion reaction provides the neutron source, it may not be as necessary to provide as fissionable material which may not be able to meltdown?

I dunno, I have no knowledge or experience at all, this is my "if I was a sci-fi writer this is a thing I would put into a story"

If done correctly you fusion fission everything to pure iron?

The light elements that you prefer for fusion, have a lower absortion rate for neutrons than the heavier elements, which you don't prefer to absorb neutrons. In other words, to get usable fusion, you need so much neutrons that a lot of material in/out of the reactor will get too radioactive.

China's actually working on a reactor like that: http://www.scmp.com/tech/science-research/article/1840219/ch...

Advantages: the fusion side doesn't have to quite hit breakeven, and on the fission side you can handily break apart any transuranic atom, even more effectively than fast reactors. It'll destroy any transuranic waste (which is most of our nuclear waste, and almost all the long-term waste), or directly fission thorium or unenriched uranium.

On the other hand, fission people don't like the complexity and fusion people don't like introducing all that hassle with fission products and decay heat. But it may well find a niche for nuclear waste elimination at least, and produce a lot of energy along the way.

Embrittlement and neutron activation are already well understood and perfectly tractable problems, since they affect the fission power industry to the same degree. That comes down to the materials you use to construct your reactor.

There are thre main problems, only one is entirely engineering the other are on the border of engineering and physics. The most pressing problem is one of plasma containment lifetime. The longer you can contain a plasma the more feasible it is to operate a fusion reactor as a power source. Right now that time is measured in seconds. An equally important problem is plasma temperature, the higher the plasma temperature the higher the fusion rate, and the farther beyond break-even the fusion reactions will be. ITER is "designed" for containment times of up to a thousand seconds, about 15 minutes, though in reality it is likely to fall far short of that. And even that is a far cry from what would make fusion power production feasible engineering wise. As it happens, tokamaks are not necessarily particularly well suited to long containment times, but there hasn't been sufficient funding to thoroughly research all of the most promising designs (such as stellarators, gas-dynamic mirror fusion, spheromaks, etc.). It seems extraordinarily unlikely that the first and most easily constructed plasma confinement system would be the most capable one as well.

The main purely engineering hurdle is cost effectiveness. Which is about being able to construct a plasma containment and heating system and all of the other components of a fusion reactor (some straightforward, some not) within a reasonable cost, not multiple billions of dollars for a single 1 GW plant. There breakthroughs in superconducting wire and so forth can have an enormous impact on the economic viability of fusion power, so it's a hugely important step.

There's also the problem of getting rid of the ash. Conventional tokamaks have divertors at the base (and sometimes the top) to get the spent plasma out of the chamber. These plates inevitably come into contact with the plasma, and we just don't have the materials to withstand that yet. We're not unimaginably far off, but it's unsolved. To add to the fun, some of the more promising materials have the rather entertaining property of being explosive under standard conditions...

Regarding brittleness - TRIZ suggested to build a reactor vessel that is replaceable in parts - or to use a shield substance that takes the damage instead. Not an machine engineer though.

Here's the paper (no paywall):

http://arxiv.org/abs/1409.3540

Interesting that the liquid blanket is the same as in the molten-salt thorium reactors (lithium beryllium fluoride (FLiBe)).

Iananp, but wouldn't that be for the same reasons? E.g. in both case you're trying to absorb energetic neutrons?

(Upon research: https://en.wikipedia.org/wiki/FLiBe#Coolant) Apparently it has properties that make it both a great coolant & a neutron moderator. Neat!

They actually have nearly opposite goals. The fusion blanket is supposed to transmute lithium to tritium [0] (create more fusion fuel), so it has capture as many neutrons as possible. The fission-reactor coolant is supposed to be transparent -- to capture as few neutrons as possible. Both of them need isotopic separation of lithium to be practical. But in opposite directions. ARC's FLiBe needs to be Li-6 enriched, and MSR needs to be Li-6 depleted.

The MSRE fuel used 99.993% Li-7 (your link), and the ARC blanket is 90% Li-6 (arXiv paper).

[0] https://www.iter.org/mach/tritiumbreeding

ARC: A compact, high-field, fusion nuclear science facility and demonstration power plant with demountable magnets

Can it be miniaturized to power a suit?

"The design could produce a reactor that would provide electricity to about 100,000 people, they say."

It would be a huge step forward if this can be built, but there still would be a long way to go. To scale that to "electricity for everyone in the USA", you would need to build about 3,000 of these (or build much bigger ones). And that's for _current_ electricity use, not if everybody starts driving an electric car.

For comparison, a 1000MW nuclear reactor produces electricity for 690,000 _households_ (http://www.nei.org/Knowledge-Center/Nuclear-Statistics/US-Nu...), and US nuclear electricity generation is about a honderd times that (same page)

A fusion reactor able to produce net electricity at all would be a massive breakthrough. I don't think the size matters much at all. Once anything is working, progress on scaling it up is likely to come quickly. The hard part is getting to the point where it works at all.

The first nuclear power plant attached to a grid was only 6MW. The first "full-scale" plant, according to Wikipedia and the BBC, had four reactors producing only 60MW each.

Honest question: how does this design scale up?

Does scaling up mean making a bigger setup with magnets of the same strength? If so, I agree it is just a matter of making the effort, but this 150-ish MW thing is "half the size of ITER", which means it already is enormous (ITER will be gargantuan with its 1400 cubic meter vacuum vessel).

Even at 'to the fourth power', I fear this would get really huge before it significantly improves on our largest fission reactors.

Alternatively, can one inject more fusion material while keeping the same magnetic field strength to get more power out without building a larger device, or would that require stronger magnetic fields? If so, are we sure we can make those stronger fields? (Correction welcome, but I don't expect we can; if we could, we likely could scale this design down)

I still think that, if this works as advertised, it will be both a huge result and only one important step on a long road ahead to 'free' clean energy.

Fusion, is one of those things that scales really well so being able to build a small power plant is a good thing.

Also, ~3,000 for the US seems like a lot. But it's only ~60 per state which is not that bad. Also, we often have multiple nuclear reactors on the same site for a range of reasons and we could do the same with this tech.

In the end we create and use crazy amounts of electricity in the US and we need any solution is going to be a large scale effort.

As long as the cost per kWh is reasonable, smaller unit size is a good thing, since it lowers capital requirements. That's why a lot of people are pushing for small modular fission reactors.

> That's why a lot of people are pushing for small modular fission reactors.

Also smaller reactors have less fuel inside and the new MSR designs don't produce plutonium or other weapons-usable byproduct, so they're less attractive targets for terrorists, dirty bomb builders and other threats, not to mention the reduced risk in failure case with the MSR technology.

I think one of the biggest issues fusion power needs to deal with is whether or not it's worth the money. Renewable energy has gained serious momentum in the past years, and while it takes up significantly more land-mass-mass (things like solar panels and wind turbines), it requires less research and development than fusion, and is advancing at a significant pace. Contrast this with something like nuclear fusion, that has been under development for decades, and still has yet to produce a long term viable product.

The one area I still think I see fusion reactors succeeding in, is that they can theoretically use nuclear waste as a fuel[0][1]. However, outside of Transatomic Power, I really don't see many fusion companies interested in dealing with that issue. Given fusions history of failure, I think that fusion energy really needs a hell of a selling point beyond a simple "we make energy" to be able to succeed. I think that selling point doesn't get much better than using nuclear waste as fuel.

[0] http://www.transatomicpower.com/

[1] https://www.youtube.com/watch?v=4UXXwWOImm8

Assuming this is the breakthrough the MIT press department makes it out to be, remaining engineering and material sciences questions can be solved and a commercially viable fusion reactor can indeed be constructed in the not too distant future, do we actually want one?

Sure, this is fantastic news in a tech-geek, star-trek sense, but here in this reality, can we actually deploy such a thing?

With the global climate in a delicate balance and energy released by fusing atomic cores or splitting them or even any process yielding energy not directly or indirectly gained recently from the sun is bound to increase the average temperature. A single reactor might not make a relevant or even measurable difference, but if a significant share of society's hunger for energy is to be satisfied this way, then it will, won't it (and if we're not sure, can we risk it)? So if this accelerates global warming, potentially even leading to a runaway greenhouse process (reducing earth albedo by shrinking the polar caps, releasing methane from the former perma-frost grounds in the Siberian tundra, etc.), then I hope they find us earth 2.0 and a way how to get there in a hurry.

The heat produced by humans is trivial compared to the heat produced by sunlight hitting the earth. That's why sources of energy that don't trap sunlight are desirable.

IANAP, but this sounds at least superficially similar to the reactor that Tokamak Energy are developing. Any nuclear physicists care to weigh in?

""" Tokamak Energy is particularly focused on Spherical Tokamaks, pioneered at Culham, because these compact devices can achieve a much higher plasma pressure for a given magnetic field than conventional tokamaks, i.e. they are more efficient.

Theoretical calculations show that a Spherical Tokamak using high fields produced by HTS magnets could be significantly smaller than other fusion machines currently proposed. For example, a compact ST power plant would have a volume up to 100 times smaller than ITER """ -- http://www.tokamakenergy.co.uk/about-us/

The lab demonstration, if and when hopefully achieved, will be great but still a long way from full scale demonstration, licensing and commercialisation. Physics may eventually work well but engineering must hold up at 1:1 and make sense from several points of view. We're not there yet, ITER in Cadarache, France, being the nearest and biggest shot. Fingers crossed.

Curious, IANAP, but suppose they get this to work? What would happen in the case of a magnetic field failure? I assume the magnetic field is generated by running electricity through the new material? What if the power fails?

They use new high temperature superconducting wires to get those higher magnetic fields. If those superconductor wires become commonplace, they have some other uses as well...

Reply on news.ycombinator.com