In the comments on my submission to the Astral Codex Ten book review contest, several people asked me about what I thought about other fusion startups. This is a summary of my responses.
Originally Written: June 2022.
Prerequisites: Book Review of THE FUTURE OF FUSION ENERGY by Jason Parisi and Justin Ball (2019).
Confidence Level: I don’t know a lot about many of their designs, especially if they are not using a tokamak or stellarator or inertial confinement fusion. Many fusion startups don’t release very much information to the public. I don’t want to make too strong of statements, but I am skeptical of all of the startups not mentioned in my review. These are quick responses to questions, not serious investigations of these companies.
I will make comparisons using the triple product, whenever it has been published. To get $Q=1$ in D-T fusion, you need a triple product of $5 \times 10^{21} \ keV \ s / m^3$. Using any other kind of fuel requires an even higher triple product. Numbers are from Fusion Energy Base unless otherwise specified.
General Fusion
General Fusion was founded in 2002 in Vancouver. They want to use hydraulic rams to compress a spinning sphere of liquid metal to compress a spheromak plasma (kind of like a smoke ring) to high density temperature. Website: https://generalfusion.com/ Requested by: Tossrock, Mark, smopecakes, & Larry Stevens.
Compressing a plasma is really hard. NIF has spent more than a decade figuring it out. That being said, General Fusion’s plasma is magnetized, so it doesn’t need nearly as much compression as NIF.
The liquid metal vortex is made of lithium-lead. If the lead gets into the plasma, it will radiate out too much energy.
I wouldn’t be surprised if some variation of the design could work. I expect that it will need several generations of experiments to get up to reactor level. They have one experiment already. I don’t know what their best triple product is, and how many orders of magnitude they have to improve it to get to breakeven.
TAE
TAE fusion was founded in southern California in 1998. They are using an accelerator beam-driven field-reversed configuration. Website: https://tae.com/ Requested by: Tossrock, smopecakes, & TheRadicalModerate.
One obvious problem is that they’re using proton-boron fuel, which means that they need to make their plasma 10 times hotter and 10 times better confined than if they used D-T.
The design is interesting and some variation of it might eventually work, but I don’t think that they’re particularly close to success. Their fourth generation experiment, C-2U (2013-2015), got a triple product of $9 \times 10^{15} \ keV \ s / m^3$, which is over 5 orders of magnitude too small for $Q=1$ in a D-T plasma and over 7 orders of magnitude too small for a p-B plasma. Their fifth generation experiment, C-2W (2016-2019), has similar density, twice as high of a temperature, and they didn’t report confinement time, so we don’t know their triple product. Their sixth generation experiment, Copernicus, should be opening soon. They still seem to have a long way to go to get to fusion.
Zap
Zap was founded in Seattle in 2017. They plan to use a Z-pinch. Website: https://www.zapenergyinc.com/ Requested by: Kyle Schiller & smopecakes.
Zap does look interesting. Z-pinches have had a lot of work done on them, so Zap doesn’t have to build everything from scratch. They haven’t been a main focus of the fusion community for a while now.
If I remember correctly, a major problem was that it is unstable to kinetic instabilities, even when it is stable to fluid instabilities. Consider a particle running along the central axis of the plasma, with a velocity almost parallel to the magnetic field. This particle will not be confined and will escape out the end. Maybe we don’t care: there aren’t very many particles like that, so we don’t lose much by losing them. But as that part of the distribution function depopulates, there is a two-stream-like instability that puts more particles in that state. So instead of just losing the few particles whose velocities were initially well aligned with the magnetic field, you get a steady leakage that ruins confinement.
I haven’t read through their papers yet, so maybe a solution is described there. It’s wonderful that they have a bunch of papers publicly available.
The velocity shear is a clever way to allow you to run more current through the plasma without triggering the kink instability. Velocity shear should also help suppress turbulent transport.
Zap is associated with University of Washington’s Fusion Z-Pinch Experiment (FuZE), which has reported a triple product of $9.9 \times 10^{17} \ keV \ s / m^3$, so they still need to make more than 3 orders of magnitude of progress.
LPP
LPP was founded in New Jersey in 2003. They are using a dense plasma focus. Website: https://www.lppfusion.com/ Requested by: ConnGator & Dan Coffman.
They are using p-B fuel, which is not the best choice for a first generation fusion reactor.
They seem to have gotten the temperatures they need for fusion, and in 2016, they reported a triple product of about $5 \times 10^{17} \ keV \ s / m^3$, 4 orders of magnitude too small for a D-T plasma and 6 orders of magnitude too small for a p-B plasma. This same report also claims that JET’s best performance is similar, which is wrong, so I’m not sure how much to trust it.
The most recent paper they published (2018) has nothing to do with fusion and instead argues that the universe is not expanding. The most recent posts on their website (July 2022) argue that cosmology is censoring their papers that “Demolish the Big Bang Hypothesis”. What?
First Light Fusion
First Light Fusion was founded near Oxford, England in 2011. They are using inertial confinement fusion with a rail gun projectile instead of lasers. Website: https://firstlightfusion.com/ Requested by: T. Oren.
Talking about the pistol shrimp makes for a cool story.
I like that they have their triple product easily accessible on their website. The “fusion measured” dot here has a triple product of $2 \times 10^{19} \ keV \ s / m^3$. They appear to have high enough density and confinement time, but 2 orders of magnitude too low temperature. This is unusual. Most techniques find it easier to get high plasma temperatures than high density & confinement times.
The biggest challenge faced by inertial confinement fusion at NIF has been in dealing with the Rayleigh-Taylor instability during collapse. To reach the best conditions, you need the pellet to collapse as spherically as possible. If the collapse isn’t spherical, then the core will start to break up before it reaches the necessary temperatures and densities. NIF has put a lot of work into making their pellets more spherical and getting the light to shine in from all directions uniformly.
Unlike lasers which shine in from all directions and especially the X-rays from the hohlraum, collisions come in from one direction. They seem to have something that makes the shockwave bend into an arc. But this doesn’t seem to me like something that can be made spherical to high enough precision. If you look at the simulations (2nd video on Targets page), it looks as though a lot of the energy in the shockwave doesn’t make it to the target and that the shockwave isn’t spherical when it hits the target.
They claim to be aiming for a shot frequency of every 30 seconds, and using larger targets than NIF. Currently, M3 can fire once every other day. More than 3 orders of magnitude improvement are needed here.
Deep Geothermal
This is not-at-all fusion, but it is an interesting alternative energy source. I am not an expert here, so don’t trust me more than anyone else. I have written about this here. Requested by: Eli Dourado, Roger Austin, & Shaeor.
I think of it more as mining the heat of the crust. It’s not renewable – they’ll extract heat a lot faster than it diffuses up from below. So the wells will run dry. But “renewable” isn’t something you should care about. What matters is if it’s environmentally friendly and if there’s a lot of it. Both appear to be true.
I don’t know if it’s technologically possible. Or rather, what depth is technologically possible and whether there’s enough heat at that depth. But it would be very exciting if it works.