In-orbit demonstration of an iodine electric propulsion system | Hacker News Reader

In-orbit demonstration of an iodine electric propulsion system

nature.com

136 pointsjollybean39 comments

This is really cool. Electric propulsion for small satellites is a hot field and has been advancing rapidly, and this is no exception. The rising cost of Xenon is a real problem in the field right now so it's great to see the work being done on alternate propellants (it would be great if we knew more about SpaceX's thrusters, which run on Krypton).

Doing some math, their total thrust efficiency seems to be around 29%, which is really good for a thruster of its size being ionized by RF power.

There’s not much to know about SpaceX’s thrusters. They’re basic Hall Effect thrusters that run on a noble gas chemically identical to Xenon but slightly less efficient (since the mass of each singly ionized ion is lower for Krypton, so for the same mass of propellant, you have to ionize more atoms than you would for Xenon… plus the ionization energy per atom may be different) and much cheaper. You could run them on Argon, if you wanted. A Noble Gas is a Noble Gas, to first and second order.

There’s no real reason we couldn’t run Xenon thrusters on Krypton except the lower density of Krypton means somewhat heavier tanks for the same mass of propellant.

Oh I'm well aware, I work on this stuff, but I'm still curious about their system, and any on-orbit data they might have. We're working on krypton Hall thrusters right now in our lab and trying to characterize the differences in operation. There's some non-obvious scaling factors we're working to understand in order to try and better optimize thrusters for krypton operation. We have a recent paper (free pdf) about this at pepl.engin.umich.edu/pdf/2021_JoAP_Su.pdf.

Assuming a spherical cow, would there be any advantage to building ion thrusters to accelerate the particles to speeds where relativistic effects become meaningful?

No, because thrust per unit power drops as exhaust velocity increases. You’ll be increasing your power supply and thruster mass far more than any propellant you save. It’s also energy inefficient to use too high of an exhaust velocity.

The mass-optimal exhaust velocity is approximately equal to the square root of (twice the thrusting time times overall power-and-propulsion system specific power times electrical efficiency).

Assuming no dry mass (spherical cow!), energy optimal Exhaust velocity is equal to the total elapsed mission velocity at any point. If you can’t adjust exhaust velocity, it’s proportional to approximately to 60-65% of the total mission delta-v (off the top of my head). More discussion of where this energy-optimal exhaust velocity comes from: https://www.youtube.com/watch?v=ogKKjpQvfuM

Indeed, one of the memorable stories from https://www.amazon.com/Ignition-Informal-History-Liquid-Prop... was when the author wrote a joke paper analyzing the optimal amount of mercury (!) to mix into rocket fuel for maximal thrust. It takes away from the power, but since mercury is heavy it slows the exhaust, and therefore can increase thrust.

To the author's surprise, the general evaluating the crazy ideas that the lab came up with failed to recognize that it was meant as a joke, and ordered that it actually be tested. So some unfortunate test site had to actually set up the test, and verify the theory.

Luckily the military came to their senses about then and never actually built real rockets using the principle.

(For those who are confused, mercury may be heavy, but is a nasty contaminant. Organic compounds including mercury have a disturbing tendency to be neural toxins. See https://en.wikipedia.org/wiki/Karen_Wetterhahn for a famous example of how little is needed to be lethal. You really don't want to be spewing it over the countryside in a flaming ball.)

Stories like this make Charlie Stross' "A Tall Tale" seem that much more believable: https://www.tor.com/2012/07/20/a-tall-tail/

Thank you so much for those links, I’ve casually been following space travel for some time and this is the first time I’ve heard about optimal exhaust velocity. I’m about 20 minutes into that presentation and it’s great!

Not really.

As the ion energy approaches relativistic, the engine also starts approaching the behaviour of a photon thruster. If your ion drive is solar powered, the system as a whole then starts to behave like a solar sail made from grey paper rather than shiny mirrors.

It would be hugely energy inefficient. Generally, you want the exhaust velocity of your rockets to be close to the total mission delta-V (or, better, the total delta-V up to that point in the mission.)

> the total delta-V up to that point in the mission.

Doesn't this give ION drives a big advantage... they can have variable exhaust velocity...

The space charge limit on the ion current increases with voltage, as does the momentum per particle. The result is that thrust/area of a given ion engine will scale as the square of the voltage. So, one doesn't want to reduce the voltage too much.

Electromagnetic rockets, like Hall thrusters, do not have the same space charge limit, and are typically operated at lower exhaust velocity.

> would there be any advantage to building ion thrusters to accelerate the particles to speeds where relativistic effects become meaningful?

Probably not; why wouldn't you just shine a light in one direction? It's effectively the same thing.

Don't photons have zero mass?

Photons have mass equal to their energy equivalent. And relativistic ions have most of their mass in the same form. So propulsion-wise, the two are almost equal...except that a light emitting diode is much simpler than a particle accelerator. Both would have very poor thrust, of course, and with existing energy sources, the waste mass from the energy production would be disproportional so neither would really be used for anything practical.

Photons have zero rest mass. That sounds like hair-splitting but it is not.

Interesting! Thanks for the paper link.

I was confused by "a noble gas chemically identical to Xenon" for a while because isn't the defining feature of a noble gas that it's chemically inert? Why not just call them Krypton thrusters?

I’d argue that being chemically inert is a property that can be compared and marked as “identical”, but that’s just semantics.

Technically they’re called Ion Thrusters. They’re just colloquially called Xenon Thrusters because xenon is the most commonly used gas to use due to its weight. Perhaps radon could work too, I’m no expert, but I imagine the radioactivity is an issue.

Usually these type are called Hall Thrusters. Ion thrusters (Gridded Ion Thrusters, the full name) work a bit differently internally than Hall Thrusters, but overall system behaves the same.

The general term is “electric propulsion,” as in Nuclear Electric Propulsion or Solar Electric Propulsion.

There's also thermal (non-ionized) electric propulsion, e.g., arcjet (which heats the propellant with an electric arc) and resistojet (which heats the propellant with a resistance heater).

Both arcjet and resistojet have been used on some satellites.

The nice thing about those is that they don't rely as much on the physical or chemical properties of the propellant. Just about anything that can be vaporized by an arc or heating element will work.

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

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

Why is it important to be a Nobel gas?

1) won’t condense on the rest of your spacecraft.

2) won’t corrode your thruster as much.

3) Chemical and fire safety aren’t concerns.

4) Is a fluid already so doesn’t need to be heated to flow.

Also, being monotomic makes it easier to model everything. A more complex molecule could have weird chemistry that ends up producing solid particles that impinge or maybe condense on your spacecraft (coating lenses, etc).

None of these are really the primary reason why we use noble gasses. It has to do with the ionization energy of the gas, which tells us how much energy we have to spend per molecule to ionize the gas.

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

However noble gases don’t have the lowest ionization energy per unit molecular mass. Bismuth, Cesium, etc do significantly better on those metrics than even Xenon does, and Xenon is better than Krypton.

http://aerospace.mtu.edu/__reports/Conference_Proceedings/20...

EDIT: and iodine, ie the subject of this study, is also significantly better than Xenon in this metric: “ A possible alternative is iodine, which is much more abundant and cheaper than xenon and can be stored unpressurized as a solid. In addition, both atomic and diatomic iodine have a lower ionization threshold, and diatomic iodine has a relative mass that is almost twice that of xenon.”

How do you gasify it if you store it as a solid? Heat?

Iodine? Keep it cold and it stays solid. But it doesn’t liquify; it is one of the elements that can sublimate from solid to gas (not sure of the physical chem here, but the transition temp to liquid vs gas from solid is so close that it doesn’t have a distinct liquid phase(?)).

So I’m imagining a solid block of iodine (it is in cold space), and when the engine needs thrust, some current from the solar cells is passed through a heating element to drive sublimation + lots of complex rocket science that I do not know of.

If you are referring to the noble gases, I don’t think they are cooled to the point of solids; just high pressure liquids..

I designed iodine thrusters for a different company. Your description of how you store/flow the iodine is correct and it's no more complicated than that. It's actually an easy propellant to work with due to its easily controlled vapor pressure (through heat) - the main issue with designing these systems is dealing with how outrageously corrosive iodine is. It. Eats. Everything.

You can technically liquefy/boil iodine (I've done it a few times). It happens around 110C IIRC.

> A Noble Gas is a Noble Gas

Probably not relevant for thrusters, but Xenon does make XeO2 and XeF6, and XeF4, and the other gases don't really do that (krypton does make KrF2)...

I was so excited by the possibility of an EmDrive engine for a bit it really seemed like everyone was trying and that there was maybe something to them. But also they would violate laws of physics so remained skeptical. Watched too much Star Trek as a kid I love to look at the skies and imagine humans travelling across the galaxy. One day.

…hot field…, hah.

The cost of things is usually a function of labour and demand, but Xenon is in sufficiently low supply that building new engines is a better option than improving Xenon farming? Where does it come from?

Xenon is a very minor trace component of air. It's collected as a byproduct during fractional cryo-distillation of air to make LOX and liquid nitrogen. [1]

[1] https://en.wikipedia.org/wiki/Xenon#Occurrence_and_productio...

> Worldwide production of xenon in 1998 was estimated at 5,000–7,000 m3

O2 annual production is 1e11kg presumably from 5e11 of air (20% of air is O2.)

Xenon occurs 1:1e7 so that air should also yield 50,000kg by mass, or 8500m3 by volume. Sounds about right. I’m guessing then O2 buyers don’t care to subsidise the Xe market.

I love that they supplied 4 citations for the opening assertion: "Propulsion is a critical subsystem of many spacecraft [1,2,3,4]". Glad I didn't have to take that on faith :)

I work on compilers and I have lines in papers like "runtime performance is important for many heavily used services [1,2,3,4]". The reason you do this is to preempt some annoying reviewer whose area is some other dimension of the same problem (eg correctness).

A company doing similar work that I shared a startup incubator with for a while is Accion. They use a liquid salt so their propellant is sort of pre-ionized.

https://accion-systems.com/our-technology/

I can’t stop laughing at the company name the authors work at: “ThrustMe”…

If anyone wants to read about EP in (a lot) more detail, here’s a great overview article: https://aip.scitation.org/doi/full/10.1063/5.0010134

I look forward to seeing more methods sections which have a step that involves launching things into orbit. I bet that at some point the exact type of rocket that is used will cause variable results due to differences in the vibrational modes modifying payloads so that their behavior changes.

Reply on news.ycombinator.com