Hacker News Reader

Something I consider to be a "red flag" in the press releases of any new kind of energy source, power plant, or engine is when they start talking about the potential applications of something fungible like electricity.

E.g.: In the Tech Crunch article linked elsewhere in this discussion there is this quote:

"Helion’s CEO speculates that its first customers may turn out to be data centers"

Do you know what else a 50MW generator could be used for? Anything. Anything that electricity is used for now. Why talk about things we all already know? Why talk about specific applications?

It's like a car company advertising that their new engine could be used to drive to Starbucks to get a coffee.

Once you notice this pattern, you'll see it everywhere in Free Energy / LENR circles...

Eventually a power source can power anything, but choosing good first customers is important. Usually, v1 of the product is expensive and flaky. So you need first customers who are willing to pay extra for the privilege of being early adopters, or for PR value. Data centers seem like a good bet here.

As you get better at mass producing the machines, the cost will come down and you can make money selling power to the grid at much lower prices.

Data centers absolutely do not go for unproven and flaky electricity sources! They go for the most boring, proven, and even downright inefficient sources. They need robustness above all things.

Untested nuclear reactors (fusion or fission) are about as far away as possible from what they want as it is possible to get.

Not to mention that many data centres are bang in the middle of high density urban areas, or at least reasonably close to them, typically in some industrial area on the outskirts of town.

No council in their right mind would approve any kind of nuclear power without years and years of environmental impact studies, justifications, reams of paperwork, etc...

You might argue that it's "safe", but the bureaucrats won't care about your opinion. They'll be worried about perception in an era where people set communication towers on fire because they want to stop 5G "radiation".

Fusion is precisely the kind of technology that is best implemented as base load in a standard power plant type configuration. Far away from big cities and scaled for efficiency.

It makes zero sense to plop something this esoteric down somewhere downtown to occasionally power a data centre during a rare power outage.

This "suggested" use case is 100% intended to appeal to people like you, the "YC News" crowd type. It is absurd on its face. Hilariously improbable. But it sounds cool and it got you talking, so a successful marketing trick, I suppose.

Google is big on solar power for its data centers, despite being intermittent. https://www.zdnet.com/article/google-1-6-million-solar-panel...

They also have grid connections, of course, so it's no problem if the energy source has frequent downtime.

Some data centers are in cities but there are many remote ones. https://en.wikipedia.org/wiki/Prineville%2C_Oregon is a small town with large Apple and Facebook centers.

Fusion reactors wouldn't be used just during outages, but for base load, reducing consumption from the grid. They already have this sort of power supply/demand matching to support their solar arrays.

Appealing to cost-insensitive early adopters is an excellent way to start. The Tesla roadster is a well-known example. I don't see the absurdity.

Solar power itself is intermittent, yes.

But it is not a risky investment – quite the opposite actually. You can buy panels and inverters with 25yr manufacturer warranties. Maintenance costs are low and predictable (cleaning panels, replacing equipment). Energy output is as predictable as the sun, and in certain regions it is pretty constant.

It is also not "costly" in the sense that the bulk of the cost of installing solar is the labor.

The sun is not intermittent

>You might argue that it's "safe", but the bureaucrats won't care about your opinion. They'll be worried about perception in an era where people set communication towers on fire because they want to stop 5G "radiation".

I agree with all the other stuff you said, and I would accept this as a fair characterization of what is in store for nuclear plants historically. I wonder though how much of that experience would transfer in the event that we have real fusion reactions, given climate considerations and a novel technology that won't necessarily become entangled in the same narratives, and (maybe?) comes with a different set of environmental implications than traditional nuclear power plants.

All practical (but not yet viable!) forms of fusion are moderately "dirty" in that they produce sufficient neutron radiation to make the reactor itself dangerously radioactive.

Compared to fission, it's much better overall, but you're still talking about remote manipulation, robots, lead shielding, etc...

There is just no way that a shipping container-sized fusion reactor will be allowed to be "plopped down" anywhere without a metric ton of paperwork and justification.

It wouldn't be safe to go anywhere near it while it is operating! It couldn't possibly have sufficient shielding in that form factor. Even if it had solid lead walls it would still be dangerous.

Fusion would be perfectly fine as base load in dedicated power plants similar to nuclear power plants. There would be thick concrete shielding, containment buildings, etc... Less than you'd need for fission, and also less waste, but there's still nuclear waste that needs to be handled in much the same way.

Doesn't Helion design require using D-He-3 fuel, which should be largely aneutronic, hence doesn't require (much) shielding?

It's not like you could plop a neutronic fuel as a substitute in a design that expects aneutronic fuel.

Where do they plan on getting the He3 fuel?

“The helium-3 is produced by D-D side reactions and is captured and reused, eliminating supply concerns. Helion has a patent on this process.” —Wikipedia

But the D-D -> He3 reaction produces a neutron. So the whole approach could hardly be called aneutronic?

"Not very neutronic".

There is actually He3 in natural helium. They can buy helium, separate out the bit they need, and sell the rest on to people who have no use for the He3.

The relative abundance of He3 is, numerically, really quite small (WP says 0.000137%, or 1 He3 per 730k He4 atoms), but that doesn't matter as much as you might think: they don't need much. Process a ton of helium, get 1.03g of He3. But it also says it is 70 to 242 parts per billion, which is a lot smaller than the other, 1370 ppb figure.

It may be cheaper to get it from used-up tritium, from people who are finished with it because it has decayed too much. In fact the US DOE does sell He3 they have extracted from tired-out stocks held ready to inject into bombs before they are sent out to use. The DOE makes its (fresh) tritium by irradiating lithium, but it starts decaying immediately, with a half-life of ~12 years, and the bombs want it fairly fresh.

These FRC reactors generate their own tritium, which is a problem, because when those fuse you get hot neutrons you don't want, and gamma rays. When you use FRC for propulsion, you can expel the tritium as reaction mass, but that doesn't work so well on the ground. On the other hand, lots of shielding is cheap on (under) the ground. But they don't make enough of it to use, and anyway who wants to bank it for years while it decays?

"Aneutronic" only applies to the primary reaction. There are side reactions, and they occur often enough that all of the nuclear safety issues still have to be handled much the same as with any other kind of fusion.

The reactor walls will still become "hot", you still need shielding, remote manipulation, etc...

It just that it takes longer for the reactor walls to reach the same level of radioactivity.

If after 1 year of operation the nuclear waste is 50% as radioactive as with a different design, that's nice and all, but it's still... nuclear waste.

See reply by csense. Also DennisP:

"They say the combined reaction will produce only 6% of its energy as neutron radiation, compared to 80% for D-T."

Neutrons are a design requirement for typical tritium breeding fusion reactors, but highly undesirable for Helion's D-He3 D-D reactor. I'd expect a 100x neutron flux reduction at activating energies over typical D-T.

You would buy a few solar panels for your house at a few thousand dollars a pop, but would you buy a 50MW fusion reactor for your house, at the prices they'd be selling at? If not, there's clearly a continuum between you and the person they'll end up selling to first.

They are rightly taking every opportunity to clarify to investors who their potential market would be.

It has to be someone without vested interests in coal supply contracts and therefore the delayed success of your product, with a huge amount of money to throw at energy security, at a large enough scale for it to be worth a big start up cost. You also need someone to go first, because fusion is scary. This is non obvious. It is an essential part of their pitch, and no amount of cringe from people who know what electricity is is worth omitting it.

If they can make 50 MW reactors at all, then yes, I will be buying the electricity from them. Not the reactor. The electricity.

It goes down wires and is distributed nationally!

I'm also not in the personal market for: Nuclear power, offshore wind, or gas turbines.

Yet, I get electricity from all of those sources.

If they can make one 50 MW power plant, then they can make ten 50 MW power plants. Put a nice little array of them on some cheap industrial land, hook them up to the grid, and start selling 500 MW like any other power plant. Easy. You can also get funding like any other power plant. Just turn up at a bank. Or issue shares. Whatever. If it works there's no need for specialised applications. It just needs to work!

There is no need to "sell" their investors on the concept of electricity generation and usage. We get it. We all get it, in the most literal sense, right now. No need to talk us into it.

They should be selling me on their capability of producing the thing in the first place, not its utility.

That's much harder if they're faking it, which is why they talk about its utility instead.

You know... if it works.

If.

Again, someone has to buy the reactor. They are in the business of selling reactors, not electricity. They are talking about who is going to buy the reactor. They think big companies who own data centres are going to buy the physical reactors. You are talking about something else entirely.

> Helion’s CEO speculates that its first customers may turn out to be data centers, which have a couple of advantages over other potential customers. Data centers are power-hungry, and often already have power infrastructure in place in order to be able to accept backup generators. In addition, they tend to be a little away from population centers.

They are definitely talking about selling them physical reactors.

It isn't as strange as you might think. Electricity isn't perfectly fungible, so new technologies do get deployed to specific use cases.

For example, solar panels got enthusiastic use in very remote areas even when they theoretically were more expensive than a grid connection. Because there was no grid in remote areas, and no population to support one.

Helion is a non-ignition fusion reactor. They aim to avoid the need for ignition by having efficient energy recapture. The website goes into more detail, and the prototypes have demoed this capability.

The lawson criterion is the point at which the plasma is heated by fusion faster than it is cooled by losses. It is a lower threshold than ignition. Even without ignition, you still need to achieve it to produce net power.

Ironically, their efficient method of energy capture actually makes their job harder than for a thermal system where heat losses due to bremsstrahlung and neutron heating are partially recovered; indeed this is the output for a conventional fusion reactor. For helion, only the energy of the plasma is harvested and thus they must exceed breakeven by enough to not just maintain but to heat the plasma by some economically useful amount.

I'm confused. I thought the Lawson criterion was the threshold for ignition, and Google seems to be backing me up. Am I missing a subtlety here?

Google is oversimplifying. The lawson criterion is defined to be the point where heat production by the plasma equals heat loss. For tokamaks (and other magnetic confinement fusion reactors where the confinement time is very long) reaching the lawson criterion is basically all you need to do to achieve ignition - once the plasma is being heated by fusion reactions faster than it is losing heat, it will very rapidly reach the point where it ignites with no further external input.

For short duration fusion (ICF, Magnetized target, Helion, etc) you can exceed the lawson criterion and thus produce gain, but the plasma may still not be long enough lived for the fusion to actually induce more fusion (which is the actual ignition point).

Thanks, I get you now.

They're not exactly looking for investment, are they? With close connections to investors with deep pockets, in a race to commercialize fusion, it's not exactly surprising that they would keep their cards close to their chest.

This is literally an article about them closing a round of investment.

What exactly would they gain from keeping this particular card close to their chest?

An advantage over competitors is my first thought. That being said, it seems sketchy to me too. You always have to take these extraordinary scientific claims with a big grain of salt

The real question to ask is, what would they gain from revealing it?

IIRC, ITER wont be fully operational until 2035, and will only ever be a research reactor. Even if it can produce net energy (I'm skeptical), it's cost and size are way up there.

Even if Helion is behind the tokamaks, perhaps this play is more about reaching an economically viable reactor design? Not first to fusion, but first to scalable fusion?

Achieving fusion is a necessary step along the road to achieving economically viable fusion. They have a long ways to go before they achieve the easier of the two steps.

>For the D-He reaction that helion intends to use, they need to achieve a similar triple product at 50 keV

The triple product is 16x higher for D-He3.[0] They also need to get D-D reactions going to produce the He3. The triple product for that is 30x the D-T value. (Yeah, you could run the D-D reaction at a loss or at barely-breaking-even I guess).

They're talking about letting the T from the D-D reaction decay to produce more He3. Tritium has a half-life of 12 years, so in steady-state, there is about 20x the annual tritium production sitting in storage. That's a massive amount -- a 1GW D-T reactor would use something like 50kg of T in a year, so that's about 1 tonne of T. (Just getting some rough approximation.) Even if you scale it down to 50 MW, that's still 50kg. It's a major radioactive hazard.

>I see no reason why a company looking for investment would hide such a result

Exactly. Fusion companies tend to trumpet their successes from the rooftop.

[0] https://en.wikipedia.org/wiki/Nuclear_fusion#Neutronicity,_c...

So basically it's a low probability win with an extremely big payoff and the added difficulty of a general lack of transparency to the public (but perhaps not to angel or series A/B/C investors), same as any startup investment.

If you don't slap some numbers on there, anything could be described as a low probability win with an extremely big payoff.

well if it goes by other startups it’s something like, given 10 startup investments…

8 crater

1 breaks even

1 does a moonshot