=flawed =energy =fusion =startups
Helion Energy is a company
working on pulsed fusion power, specifically with magnetic confinement using
a field-reversed
configuration
("FRC"). Interest in those declined in the 1970s because tokamaks had better
performance - that's the "conventional" approach used by
ITER.
Helion does not have a
way to get lots more fusion from FRC reactors. If they did there would be
signs I'd notice, and you'd notice too because they'd be bragging about it.
Instead, their plan is to recover most of the energy put in, so you only
need a little bit of fusion. Let me explain why that doesn't work.
The easiest kind of fusion is D-T
fusion, but
80% of the energy is released as fast neutrons, which obviously can't be
contained magnetically. Helion wants to recover energy electrically for
efficiency and cost reasons, so they instead want to do D-D fusion, and also
D-He3 with He3 made from the D-D fusion. D-D is about 1/30th as fast as D-T
under good conditions for fusion, and D-He3 is much harder than that since
He3 has twice the charge.
Current tokamaks do not produce net power
with D-T. The record is ~0.7 for fusion power / input power ("Q"), and
conversion is inefficient. Helion's design produces less fusion per input
power than tokamaks, with the same fuel. D-D produces much less energy than
D-T. Therefore, Helion is not close to making power.
Helion has
fans, who will reply with something like: "But Helion can recapture most of
the energy, so the Q needed is much lower! They demonstrated 95% energy
recovery!"
And if that's what you're thinking, congrats, you fell for
their deliberate ambiguity.
It's very easy to get 99% efficiency
with a transformer or inductor; you put energy in a magnetic field, you turn
it back to electricity, it works well. When Helion talks about 95% or
whatever energy recovery, they're talking about energy put into a big
electromagnet and then taken back out - not energy put into a plasma and
taken out of the plasma. And when it comes to fusion Q factors, the relevant
number is the energy put into the plasma, which in Helion's
case is much smaller.
So can they get that 95% recovery for energy
put into their plasma? No, that's physically impossible, because:
- With D-D
fusion, radiative losses (to X-rays) are always >1/3 the fusion energy
produced. (Unless you have something big enough to be "optically thick",
like a star.)
- Some of the particles escape the magnetic confinement.
With a FRC design (which makes fairly unstable plasma states) this is a
problem even with short pulses.
- Extracting almost all of the energy
from a very hot gas requires expanding it a lot. The effective expansion
ratio of the plasma with something like Helion's design is limited, so
there's a lot of residual heat.
other fusion approaches
If
I don't like Helion, is there some other approach to fusion power that I
think is more practical? (Not counting "using sunlight" as "fusion power",
that is.) Sure, getting net power from fusion is easy: just set off nukes in
an underground cavern full of molten lead or something, and then use that
heat to generate power. That's not even close to economically competitive,
but it's entirely feasible.
How about controlled fusion without
explosions? Tokamaks are more expensive than fission power (which is
expensive) but would they work? My view is that plasma stability inevitably
decreases with fusion power density, and makes pure magnetic confinement
unworkable for economically-useful fusion power levels. But that's a much
more complex argument that I have less confidence in.
why do I care?
Isn't it
fine if investors lose money on some speculative investments because they're
not technical experts on everything? Why is there a need to criticize
startups like Helion?
Money is wasted on lots of stuff, and I don't
care much about the money Helion is wasting. (I don't think Commonwealth
Fusion will be successful either, and they raised a comparable amount of
money, but I'm not writing a post about them.) What bothers me more is that:
1) Sam
Altman managed to use a
story about "making fusion happen Real Soon Now" as a means to get support
from people like Paul Graham and gullible nerds at OpenAI.
2) Helion
takes up cultural space that could go to something better - because they (or
some investors) have been paying for PR pieces in magazines and newspapers
and youtube channels.
3) I'm not going to support a company as "maybe
advancing humanity" when they refuse to publish any actual data because it
looks bad.