Compact electron accelerator reaches new speeds with nothing but light | Hacker News Reader

Compact electron accelerator reaches new speeds with nothing but light

phys.org

113 pointsPaulHoule54 comments

Thought Emporium made a really good video about this kind of accelerator a few years back: https://www.youtube.com/watch?v=UZbXGDxMRCw

Can this be used to generate soft X-rays, aka "extreme ultraviolet"? You can do that with a synchrotron or a big linear accelerator, but it hasn't been cost effective. A smaller accelerator could replace the vaporized-tin hack ASML uses as a light source.

It's a bit much to call it a "hack". ASML & collaborators have created an EUV source that works at the scale needed to mass manufacture chips. It's not something held together with duct tape. Yeah maybe a little accelerator is theoretically elegant, but some among the thousands of engineers and scientists that have worked on their EUV probably considered that.

To add to this, here is a great video on some of the technical/physical challenges ASML had to solve for their EUV source: https://www.youtube.com/watch?v=5Ge2RcvDlgw

That's one of the applications, yes, compact x-ray source with smaller source size.

Actually generating x-rays is easy. Every dentist office has a "high-luminosity" x-ray generator

But x-rays are too short-wave for EUV. Maybe they have considered some form of down-conversion from x-rays for EUV but at those wavelengths it's not very trivial.

To generate soft x-rays/UV you don't really need an accelerator, since high voltage (not that high actually, since required kilovolts can easily be achieved) is enough to produce photons of those energies.

Vaporization you mention is likely because of electrons hitting the material, since in any case to emit photons electrons need to be slowed down. Either electrical potential difference in a vacuum chamber or an actual accelerator is being used, the material to brake electrons will evaporate anyway.

People have built free electron lasers to make EUV. I’ve seen them for sale.

Whenever I see a scheme like this, I want to ask "what's the quality of the beam?" Particle physics experiments, particularly colliders, need beams that are not just energetic, but also extremely bright, so the particles can be focused down to tiny spots to get sufficiently high collision rates. The particle energies must also be within a narrow range, the narrower the better.

The proton bunches in the LHC are just 2.5 microns wide. They have to be to get the luminosity high enough.

Yep.

It is true for colliders and also a big issue for heavy ion fusion, which on one hand seems more feasible than laser fusion because of the very low efficiency of laser drivers in comparison, but boy you need a huge (kilometer scale) machine and probably 100 or so independent beams that are very high quality and well synchronized.

A major problem with laser fusion is getting the final optics to survive. The designs I've seen put grazing incidence mirrors at considerable distance from the target, causing the entire thing to be large, much larger than the containment building of today's PWR power plants.

https://fti.neep.wisc.edu/fti.neep.wisc.edu/aries/JUN2001/op...

Heavy ions, being charged, can be deflected by magnetic fields, so nothing need be in line of sight of the target.

I don't think this kind of technology will ever create the kind of beans particle physics use. There is a hard limitation, because once the particles move fast enough, the only known source of the "light" needed to accelerate them further is particle accelerators.

Instead, it looks like something laser focused on materials science and engineering. Think about what you can make with a $10k tabletop cyclotron light source.

Bad ass! Our E&M professor was kind of flabbergasted that we weren't doing table top colliders (this was a decade ago), almost implying some sort of scam was going on.

Do most people know that alchemy has been realized in modern physics? Lead has indeed been converted to gold (at terrible efficiency)!

The alchemy moment will be when a laser impulse (made of a cardboard, a stick and a couple mirrors) assumes the shape of the gold nucleus.

> Lead has indeed been converted to gold

Seaborg did it in 1980 but with bismuth.

“It would cost more than one quadrillion dollars per ounce to produce gold this way.” And that’s in 1980 dollars :)

https://youtu.be/dpBwlcUZZRI

At least we know we can come up with more gold if we deplete the planet one day.

But only Uncle Scrooge will be able to afford it.

Money is an artificial construct. If it’s worth doing, it will be affordable to those who get value out of it.

Money is made up, but cost is real. It's not the dollar value that matters here.

Ultimately it's energy. And as long as the output is valued enough, we would put the energy there. It would be a terribly inefficient/poor world but it could be done. Unless we have so much energy available that the cost is negligible.

Energy is finite, and without some form of subsidy this kind of use case has to compete directly with everything else that we use energy for. Production of usable energy is not free...

This is a very optimistic take on the efficient market hypothesis, and not one that's well supported by the evidence IMO.

This is like saying that miles are an artificial construct. The problem is that distance is not.

Or in pop-Zen terms, don’t mistake the finger for the moon.

If you want the lowdown from an expert, the issue with these acceleration schemes with lasers is the accelerated beams aren't that great in terms of characteristics, specifically they tend to not have the best divergence and energy spread compared to traditional linacs (divergence is better for electron beams from LWFA but for the ion accelerators which are another major application of intense laser interaction with plasma, the beam characteristics are much worse). With a cyclotron or a linac you control everything a lot better because the involved fields are fields you put in yourself with equipment you can tinker with, but here the plasma bubble is microscopic and the relevant physics happens on a femtosecond-to-picosecond time scale, and any small issues with the input laser (specifically the wings of the gaussian pulse) mess things up.

One of the issues that I sort of sniffed out in the field is as a whole they are pushing for higher and higher peak energy without really looking to improve beam quality, that is, everything below the peak. Like this paper (which an old colleague is an author on!! Congrats to him) is actually pretty good on the divergence as 1mrad (mili radians) but the energy spread is 15% around the peak, which is good for some things, but there are experiments with linacs that require percent spreads less than a percent. Now, you can probably "filter" out the energies you don't want out of the beam to make it even more monoenergetic, but for example this beam has tens of picocoloumbs. Assuming (very generously, haven't read beyond the abstract yet) that most of the tens of pico they report is in that 15% spread, 1% means you might be dipping into less than a picocoloumb territory on target for your experiment, which might not give enough dose for what you need to probe or otherwise do.

Anyway, in terms of the field, this work is fantastic, and generally Howard and Jorge are fantastic scientists (never met Jorge but Howard's a real decent man), but this is the cream of the crop of what we as a field can currently do. I feel like getting to the point where we give the traditional accelerator types a run on their money is still a few years out.

Thanks for the input. I've been hearing about wakefield accelerators for a long time, but never got into the nitty gritty of it.

Is the beam quality a fundamental limiting factor, or is it solvable (at least theoretically).

It sees they're going for the get the energy up and solve the beam characteristics later approach

caution: Haven't studied particle accelerator design

My understanding of most particle accelerators is that they accelerate their subjects directly with fields, either oscillating or steady. Wakefield accelerators use a combination of fields (the laser pulse) and other particles (the "stationary" ions). My suspicion is that the ion background introduces enough noise/chaos into the process that it would be a great effort to make the output more monoenergetic and retain total brightness. But I dunno, I always found particle physics a little tedious and didn't pay as much attention as I could have.

That is also true, basically target imperfections plus noise in the laser cause imperfection in the acceleration process. Essentially the shape of the bubble can change the accelerating fields uniformity, leading to the divergence and spread issues. They usually cryo their target to minimize just plain chaotic behavior from it being a gas, so generally most of the jitter is due to wings of the laser itself causing perturbations in the gas/plasma (it's a plasma once the peak reaches it). That said, you're totally right about the target itself being a source of noise too.

> Is the beam quality a fundamental limiting factor, or is it solvable (at least theoretically).

At least frequency noise is inherent for any accelerator where the particles stay for a very short time.

Everything is solvable theoretically :) I don't think it's necessary fundamental limitation to the process, but it seems a bit far away just at this moment. I will be a little more charitable to myself and the field generally, some people do care about characteristics, it's just what gets the phys revs+ seem to be higher peak energy. It's just a shame because for applications the general beam quality is a potential show stopper. I think the problem is that as anyone who works with numbers knows, including hackers, the peak value of something is what you put on the slide because it's easy and sounds sexy and is more exciting than a histogram.

The first pulse that creates the electron channel sounds similar to the army's lightning gun:

> The Laser-Induced Plasma Channel (LIPC) is designed to hit targets that conduct electricity better than the air or ground that surrounds them.

https://www.bbc.com/news/technology-18630622

Now that is some proper sci-fi weaponry. Questionable if practical, but so cool

Tesla tank when?

I wonder if this could work for protons - because if we could accelerate protons into Boron-11 we could have aneutronic fusion which would generate unlimited clean energy, even cleaner than the current fusion designs.

you have to produce the photons as well, the particle flux is likely extremely small with those devices so you may need a gigantic power input. Remember that a proton is almost 2,000 times heavier than an electron.

You don't need a wakefield accelerator to get protons up to the energy needed for p-B11 reactions, since the peak reaction cross section is only around 600,000 eV (as opposed to the 5,000,000,000 eV in the article). You can already do this with an accelerator which would fit in a room.

More important parameters for an accelerator of that purpose are efficiency (wakefield accelerators have incredibly poor efficiency) and brightness (how close you can get the particles together, both in space and in velocity).

The real challenge for beam-based fusion isn't getting an accelerator for your beam, it's making the reaction happen where there aren't a bunch of cold electrons present to absorb your beam's energy before the fusion happens. This usually turns the situation back into a plasma containment problem.

A proton is around 2000 times as heavy as an electron, so your energy numbers are actually semi equivalent. (At least that's how it seems to me - I don't know if this would even work on protons.)

I hear your point about efficiency though. Why do you need similar velocity for the particles? It seems to me you just need a LOT of particles, but beam quality doesn't matter much.

The cold electrons you mention are the ones bound to the Boron? Can we give the Boron a positive charge? All the charge will live on the surface, which is exactly where we want it.

Make the box metal to the height of the boron, then an insulator after that, and charge the box. Put the output of the accelerator directly above the boron (the output is facing downwards), and keep pumping out helium.

Cool the box from the outside to capture energy, and also capture electrical current from neutralizing the charged helium.

Am I missing something major?

> Am I missing something major?

Multiple things, I'm afraid.

I'm answering this as more of an amateur fusion enthusiast than a physicist, so I expect this list is a rough guide not a complete and precise one:

• A wakefield accelerator has a plasma region significantly more massive (in aggregate) than the particles being accelerated, and the acceleration comes from the electrons in the plasma being easier to move than the ions, which creates a high charge gradient, and that charge gradient accelerates the particle beam.

The plasma is "cold" compared to the beam.

• You give boron a positive charge by removing electrons, but the more you remove, the harder it is for them to fuse (they mutually repel before they get close)

• Even when the nuclei hit perfectly and in optimal condition, they mostly just bounce off instead of fusing.

This releases photons (because all acceleration of charges always does that), which saps energy away. Stars only get around this by the photons being reabsorbed almost immediately and re-heating the plasma in the surrounding volume.

• While the setup you describe will do fusion if you pick the right energy level, the energy level of this accelerator will instead mostly produce heat, tau-antitau particle pairs, and possibly some proton-antiproton (and neutron-antineutron?) pairs.

• For stuff like this, the interesting thing is the energy not the speed, so while (in Newtonian physics) 5 GeV electrons go at the same speed as 2.5 MeV protons or 227 KeV B-11, that's wildly into the relativistic range for electrons so they don't do that anyway.

You wouldn't use a beam for fusion anyway, somewhat because of the "cold electrons" but you want it to be in a randomized soup that you can then compress, which is why ICF is a better scheme for fusion.

It won't. We as a field do accelerate protons by via another means, typically by hitting solids instead of gas jets like the article. And...would you know, we have done pB-fusion in this way :) [0] It's nowhere near efficient to be a "net energy gain" (in fact, I'm not sure it can be by just using the ion acceleration schemes we have), but these fusion experiments with lasers are useful as compact spallation sources.

That to me is even more exciting because as others pointed, the lightest ion (proton) is 1836 times the weight of an electron meaning current ion accelerators and neutron sources are large facilities vs. electron accelerators which need not be so large (doctor xray for example). Minimizing those could be a game changer, even if the beam characteristics aren't that great yet or the efficiency is bad (which is the issue for laser acceleration in general).

[0] https://journals.aps.org/pre/abstract/10.1103/PhysRevE.101.0...

This could be used for Directed Energy Weapons? Strategic Defense Initiative seems to be getting traction lately.

Could this be used to make some variant of an ion drive?

No, because the total accelerated charge is tiny as of now, so you don't get a lot of total thrust. You'd need to scale up which hasn't really been considered yet. I'm not sure if the plasma bubble will be stable if it is larger (which I think is one way to up the total charge).

Total thrust matters a lot less than how efficient it is.

Could? Sure.

Should? Nope.

Ion drives are an excellent way to make very little reaction mass go a long way. While something like this could do that, for very high energy particles you rapidly get relativistic effects, and the ultimate relativistic effect — zero reaction mass having finite momentum — is from using light. This can be as simple as a warm bit of the vehicle[0], but if that heat energy comes from solar then you might as well make it even simpler by replacing the PV with a mirror and suddenly your space ship is propelled by a solar sail.

But you could use this if you really wanted.

[0] the most recent explanation I've heard of for the Pioneer anomaly

The general problem with ion engines is that the exhaust velocity is too high. This makes them energy inefficient. Ideally, you want the exhaust velocity of a rocket to start out very low, then increase as the vehicle goes faster. That way, the ejected reaction mass is left stationary in the initial reference frame, with no kinetic energy (or, nearly stationary, with low kinetic energy). All the energy therefore goes into the kinetic energy of the vehicle.

This is also part of why using liquid hydrogen in the first stage of launchers is dumb, vs. using hydrocarbons.

Given all frames of motion are relative, I thought it was more about gravitational potential than kinetic energy?

Any frame of reference can be chosen, and energy is conserved in each. So choose the reference frame in which you are initially at rest.

Hmm. So maybe using this to accelerate protons would be better.

(I also wonder what happens if a spaceship bleeds a measurable fraction of its weight in electrons, then tries to coast away from the resulting negatively charged cloud.)

0.05% of your total mass is electrons[0], so you can't do that even if you removed all of them.

Well before you even get that far, your electric field strength exceeds the binding forces of the molecules you're made of, so you explode.

Well before you get to the point of exploding, your net electric charge is too strong for you to eject more electrons, and any electrons in the ambient environment will fly towards you, producing a shower of deadly radiation when they hit your hull.

[0] assuming you're mostly hydrogen; a bit less than half that if you're mostly made of heavier elements like Uranium

For designing relativistic rocket engines, there's no point in using lasers to accelerate ions to close to the speed of light. You can just use the laser directly -- the other stuff is extraneous. A relativistic electron has no more momentum than a photon of the same energy.

https://en.wikipedia.org/wiki/Photon_rocket

The photon rocket is the most widespread type of rocket these days, eclipsing both ion thrusters and traditional chemical rockets.

...flashlights ;)

As an indirect mechanism for triggering more powerful reactions that could be used as the primary drive, I think so.

My dream spaceship has a fusion engine and VASIMR drive, a kind of variable ion thruster, for variable thrust from takeoff to ~.1c speeds.

But fusion reactions are hard to ignite and sustain. Putting a tokamak in a spaceship might be possible, but they weren't really designed to scale down.

Fusion reactions require energies in the MeV range to ignite, but that's with perfect conversion. I'd like the flexibility up to GeV range, or many small 100 MeV units in parallel like this accelerator or the BELLA & kBELLA at Berkeley[kBELLA].

These can also be chained. Cascading wavefront accelerators appear to be the most compact way to achieve scale for high energy particle beams, perhaps combined with lattice-confinement-based fuels[lc-fusion]. Berkeley is also developing software to model "multi-TeV high-energy physics colliders based on tens to thousands of plasma-based accelerator stages."[warpx]

Cuz your spaceship needs a laser gun too, right?

[kBELLA] https://newscenter.lbl.gov/2021/12/06/fiber-lasers-poised-to...

[lc-fusion] https://www1.grc.nasa.gov/space/science/lattice-confinement-...

[warpx] https://www.exascaleproject.org/research-project/warpx/

> Putting a tokamak in a spaceship might be possible, but they weren't really designed to scale down.

It would be senseless. I mean, ITER, if it could run continuously (it can't; it will run a few weeks before materials reach radiation limits, and it can't breed tritium) would take 300,000 years to fuse its own mass in fuel. Smaller tokamaks are not much better.

There's no application in space for controlled DT fusion where fission wouldn't be superior.

You would do better with a proton accelerator.

Reply on news.ycombinator.com