Hacker News Reader

Comment by Game_Ender | Hacker News Reader

Lots of interesting fusion startups. This group is using a gun type design that reminds of the Fat Man atomic bomb [0]. Except here it's a fusion target hit by high speed slug causing is to rapidly compression and undergo fusion. The key things is that unlike the NIF they have a clear path to power extraction. In production they are planning to use a chamber with circular sheets of falling liquid lithium to capture the fusion neutrons then transfer the heat [1]. Breeding some tritium along the way.

0 - https://en.wikipedia.org/wiki/Gun-type_fission_weapon

1 - https://firstlightfusion.com/technology/power-plan

View on HN

Cool idea. Capital costs would certainly be lower than any of the magnetic confinement designs, if it works.

The power plant design they show has a 150MWe target power. Will be an interesting engineering challenge scaling it up and keeping all the finicky little parts and seals in the gun working when by design it's connected to a (small) nuclear explosion by a long pipe. If the timing is right you could have a heavy rotating shutter shielding the muzzle from the backblast.

They will, fortunately, not need to solve any of those problems, because there won't be any power plant. Instead, they will spend all the investors' money and then go do something else.

What's your take on Helion? They're acting exactly like every other scammer on the surface, but the more I dig the more I find evidence of thinking about things like how much quartz will evaporate and get into parts you don't want it to over 10 years. I can't see an overriding reason why it's immensely stupid like all the other schemes out there either.

For a while they were describing D-He3 as "aneutronic", (that reaction is, but a mixed D-He3 plasma will be producing D-D side reactions, which produce neutrons) and some of their promotional materials showed reactors in standard ISO shipping containers. (missing the concrete biological shield it would need) However, I just checked their site again to link to those mistakes, and they've been removed, which is nice. Their CEO also made some nonsense remarks about how Helion reactors would be deployed directly to end users like datacenters, which the NRC wouldn't like.

They're a lot like the other fusion startups in the current scene. They have a machine, about which they make various surprising claims, but keep almost all the specifications secret, since it's a privately held venture rather than a public government-funded experiment. It's not obvious incoherent nonsense like solar roadways from a couple years back. Maybe it will work, maybe it won't. Check back in 10 years.

The D-He3 reaction itself is aneutronic. Overall, Helion's reactor would produce only 6% of its energy as neutron radiation, compared to 80% for D-T reactors. For Helion that means they extract electricity directly from fast-moving charged particles instead of using a heat cycle. In practical terms that's often what people are thinking about what they refer to aneutronic fusion.

The NRC and their UK equivalent have already been moving towards lighter regulation for fusion reactors, and that's for D-T reactors. A reactor like Helion's could well be regulated more like medical devices than fission reactors.

I'd lump all the other fusion startups into the same category as incoherent nonsense like solar roadways as they all rely on recycling energy through a steam turbine (which is already prohibitive even if you had free heat) in order to power their rube goldberg machine and somehow that's supposed to be 100x cheaper than existing steam turbines when excluding the fuel.

I was wondering if there was some incoherent nonsense I'd missed.

Helion is the only fusion startup I know of that is not obviously a scam.

They count on D-D side reactions for neutrons to use breeding tritium, which decays over decades to the 3He they need. I don't see how that could produce enough. If not, their thing might still end up usable to power outer solar system space probes, using 3He decayed from the usual CANDU-produced tritium.

The Princeton FRC design is intended for spacecraft propulsion. I don't know if it is more or less practical.

They say they'll use pure lithium but I don't see how that can work. One lithium nucleus plus one neutron makes one tritium nucleus. If you lose any tritium at all, you run out after a while. Other fusion projects include lead or beryllium, which releases two neutrons after getting hit by one. That keeps their tritium breeding over unity.

On top of that, First Light says they'll use natural lithium without enriching it to lithium-6, which makes things even worse.

If they used 50% enriched 6Li/7Li, they would breed extra tritium. But there would be no point, because they would have no way to refine it out. They don't care, because they know there will never be an actual power plant. They will spend investors' money, have some fun, and move on.

  > But there would be no point, because they would have no way to refine it out

Why not? I'm not a subject manner expert, but from my armchair it seems that refining one element from another is far more straightforward than refining one isotope from another as is done with Uranium.

I would expect that hydrogen and lithium should be able to be refined chemically, as both are very reactive elements. And even if mechanical refining is necessary, we have much industry experience with that: Uranium isotopes have been mechanically refined for almost a century. Furthermore, lithium and hydrogen are far more dissimilar than are U235 and U238 - lithium has six times the atomic weight of standard hydrogen and double the atomic weight of tritium.

See, this is how they find pigeons to hand over all their money.

"Gee that guy made a good point...I don't really have an answer...I know, call him a pigeon!"

Heavy water reactors have been producing tritium at scale for decades. India is talking about deploying a bunch of new PHWRs. They're recovering tritium from water, obviously. Would it be much harder to recover it from molten lithium? (Lithium hydride complexes?)

Seems like a hydrogen gas in a vat of liquid metal or salt is just going to just bubble right out.

At PPB concentration?

Literally anything can seem, to somebody.

I guess literally anything can be "PPB concentration" if you don't care how many parts it is per billion.

A 1GW fusion reactor would consume about a ton of fuel per year.[1] That's about 600 kilograms of tritium. Over the course of the year a reactor with a positive breeding ratio will produce more than that.

The breeding blanket for ITER will be about 2000 tons[2] and ITER is designed to produce 500MW.[3]

This means that ITER will produce at least 300 kg tritium in 2000 tons of blanket, over the course of a year. Collect the tritium annually and you've got a concentration level of 150 parts per million, or 150,000 PPB.

[1] https://ccfe.ukaea.uk/fusion-energy/fusion-in-brief/

[2] https://www.iter.org/mach/Blanket

[3] https://www.iter.org/proj/inafewlines

You need to get that tritium out every day to have something to operate the reactor on. Look up the worldwide stock of tritium sometime.

Doesn't mean you have to do it every day. CFS ARC is a 270MW reactor. In a month it will consume 13.5kg tritium. Current world tritium supply is 20kg, though we could produce more from fission reactors if we really wanted to. 13.5kg out of 2000 tons is 7 parts per million, but with a breeding ratio over one, more than that would be produced. Tritium inventory to start up the ARC reactor is 100 grams.

Reply on news.ycombinator.com