dark glasses, game faces.

fusion happens when atoms hit each other really hard, but they dont usually want to hit each other. they need to be forced into it. you can do this by getting them really hot, which makes the atoms move faster, which much like cars doing 90mph in a 65 zone increases the chance of collision

really hot atoms in a really hot gas become what's called a plasma, where all their electrons fall off so the leftover positive nuclei are floating around with negative electrons

therefore, fusion research consists of trying to get plasma hot enough to fuse.

two problems

it has to be really really really hot to fuse

really really really hot plasma is bad for nuclear reactors, humans, and everything that isnt the sun

solutions arise

plasma is electrically charged, so we can control it with magnets. magnetism and electricity happen at right angles to each other, like if an electron is moving vertically a magnet can pull it horizontally, and vice versa

if we get the plasma to whirl around in a circle, we can hold that moving plasma flow in place with magnets

behold! the tokamak

it would be nice if it worked out that neatly, but plasma is really messy and ends up drifting all over that place. it turns out that doing the spiraling helical thing in the diagram is the best way to contain it, but it's not 100%

and even trying to confine it like that needs a huge amount of energy to get a huge amount of electricity flowing through the plasma to get it magnetically active enough for the magnets to hold it in

when people speak of "nuclear fusion in ten years!" they mean that we will achieve "breakeven", when the amount of energy used to just run the thing is "worth it" because we get more energy out

but plasma remains messy, and every time it messes up it costs energy and damages the reactor. an early tokamak had a hole melted in it by what we later understood to be very similar to a literal lightning bolt

the tokamak is a somewhat brute-force approach. what if there was a more elegant method

though we don't yet understand all the messes of plasmadynamics, we understand many of the biggest ones, so we can come up with ways to cancel them out.

one of the easiest ways is rotation: if the plasma drifts left, let it get halfway and then flip it upside down so it drifts back right. or constantly keep rotating it. you can see the same idea on the helical magnetic field in the tokamak diagram from before

so in a stellarator, we've switched our focus: instead of relying on flowing plasma current to respond to the magnets holding it in place, we've reshaped the magnets to catch plasma drifting outwards and redirect it back inwards.

pros: no messy, power-hungry plasma current to force. a stellarator can operate in "steady-state": instead of constantly touching it up like pottery on the wheel, it just goes until you turn it off (or it runs out of fuel or explodes of whatever). since we're already fucking with the magnet shapes, we can also put in extra little twists to trap plasma messes, smooth out the more obvious tangles, etc.

cons: all that magnet fuckery is complicated. it's harder to calculate and design, harder to physically build, and harder to maintain. depending on the design you can make it so the magnets are movable, so for a research stellarator you can try new magnet designs a little easier, for example, but that movable magnet will never be as good as a properly fixed magnet. also the magnets have to be quite close together so there's not really space to build, like, a door into the interior for maintenance or access

and the real con: tokamaks are better. they quickly got much closer to breakeven as we just continued to refine our plasmadynamics alongside brute-forcing the plasma current

BUT

stellarators are getting better

i almost wrote that theyre getting beautiful. they are. just look at her