Throwing any old rubbish into the propellant tank with an electric engine

[SomeGuy12]

Out of the available electric engine designs, why is Xenon gas such a popular propellant? It's a little hard to come by.

Second, are some of these designs more fuel-agnostic than others? An ideal engine would be one that is high thrust, high ISP, lightweight, long burn life, and it would let you throw any old rocks you found out in space into the input hopper.

I guess you'd convert the rocks to plasma first, then separate them by atomic weight. Then with this atomic weight-separated propellant, you'd set your magnets or electric fields or voltages or whatever for a propellant with that particular atomic weight, and use it until you run out, then switch to the next one.

Basic physics say it shouldn't matter what the propellant is, just what your desired ISP is...

[spudman2]

What you've said sounds correct, but rock-vaporizing-separating equipment is heavy and easily broken - two things spacecraft do not want to be. I've heard a similar case for NTRs, and while maybe not any old rock can be used as propellant, I see no reason why water, or CO2, or any fluid you just... find couldn't be used in an electrical engine. Even more so for NTRs.

[KSK]

I believe xenon is used because its a noble (so doesn't corrode your electrodes) gas thats easy to ionise. Most of the energy used to run an ion drive goes into ionising the propellant.

[Streetwind]

Pretty much what KS says.

A plasma of material with that much energy pumped into it is extremely aggressive when you concentrate and accelerate it through a magnetic field. You want it to be a chemical element that is so stable that it doesn't want to be aggressive, not a random mix of elements that will happily react with everything it touches.

Also, xenon is a very large element. The largest stable noble gas, in fact. That means it has a huge number of electrons in its shell, and stealing one to ionize it is very easy. By contrast, a small element like helium has only two electrons. Stealing one equals to taking 50% of what there is to take, and the atom is going to fight this process hard. It takes much, much more energy to ionize a small element than it takes to ionize xenon. So if your energy budget is finite, using xenon means you can spend the largest part of your energy budget on accelerating the propellant, thereby getting the best efficiency for the energy invested.

The problem with proposals like the rock vaporization thing is always that of efficiency. Sure, xenon is expensive. But your rock vaporizer is going to be relatively big, heavy, and take an absurd amount of power just to prepare propellant for consumption. You could just as well toss the rock vaporizer and feed all of that power direction into a classical electric engine with xenon propellant and get better thrust, better Isp, less upmass to be placed into orbit and a simpler, longer-lived and more reliable vehicle overall.

As far as fuel-agnostic electric engines go, there are a bunch that can theoretically eat almost anything. Just nobody does it because using xenon is simply better, and when you're dropping three-digit millions on construction and launch, the price of a few hundred kgs of xenon starts to be pretty insignificant.

Then there are proposals like the VASIMR, which is supposed to run on argon (another noble gas). Argon is flat-out worse than xenon in every respect, but it's much more common and thus much cheaper. And since the production model VASIMR is expected to start at 200kW and scale up from there - which is orders of magnitude higher than any electric engine in space today - fuel consumption is going to be higher and the company thinks that it starts making sense paying attention to the price tag of the fuel.

[Woksaus]

I'm currently working in a lab where they did research on using metal propellants for Hall thrusters. It's promising because the propellant takes up a lot less space (metal pellets as opposed to a big gas tank) and is much cheaper than Xenon. Plus it gives you all kinds of pretty colours, depending on the metal that is used

Sources:

http://www.me.mtu.edu/researchAreas/isp/

http://arc.aiaa.org/doi/abs/10.2514/1.47410?journalCode=jpp

[PB666]

Out of the available electric engine designs, why is Xenon gas such a popular propellant? It's a little hard to come by.

Second, are some of these designs more fuel-agnostic than others? An ideal engine would be one that is high thrust, high ISP, lightweight, long burn life, and it would let you throw any old rocks you found out in space into the input hopper.

I guess you'd convert the rocks to plasma first, then separate them by atomic weight. Then with this atomic weight-separated propellant, you'd set your magnets or electric fields or voltages or whatever for a propellant with that particular atomic weight, and use it until you run out, then switch to the next one.

Basic physics say it shouldn't matter what the propellant is, just what your desired ISP is...

The Vasmir is set to run on Argon, it will be the most powerful ION drive out there when completed. The reason that argon is

Momentum equation = to get the most delta-V on a ship you want to have the most mass on the accelarant.

Weight density per volume. You can store more of that mass in a unit volume with argon.

Most controllable and least damaging Ion. The ions are going to tear the crap out of the accelerator grids, the more that miss their mark, the more damage they do.

With electric and hard masses you can use a high velocity rail gun. But VASMIR once it gets going is going to surplant other types of ION drives. Unlike other drives it does not rely on plates, it stirs the plasma into a gyre and focuses the gyre using microwaves so that it basically does not have the grid distruction. Just like Cannae drive, it needs to get its legs in space.

[impwarhamer]

when your talking about the costs of launching fuel into space, having a fuel thats a bit more pricey isn't that much of an issue.

[SomeGuy12]

when your talking about the costs of launching fuel into space, having a fuel thats a bit more pricey isn't that much of an issue.

No, but you're out at Ceres, having landed on it. You'd like to be able to stuff whatever you find lying about into your nuclear-electric rocket to improve your delta-V budget for getting home (or to the next asteroid)

It might be supplemental at first - you'd have the minimum amount of fuel to get back, but if you can find more, you can get back on a faster trajectory. Eventually it would be part of the mission plan to pick up fuel at the destination.

[SuperFastJellyfish]

The ELF thruster is supposed to be able to operate on many different propellants. It is still in early testing though, AFAIK.

http://adsabs.harvard.edu/abs/2009APS..DPPCO7001W

The Electrodeless Lorentz Force (ELF) thruster is a novel plasma thruster under development at MSNW and the University of Washington which utilizes Rotating Magnetic Field (RMF) current drive technology to ionize a neutral gas and drive an azimuthal current to form a Field Reversed Configuration (FRC) plasmoid in a diverging magnetic field. The magnetic gradient imparts a net force to the FRC which is ejected from the thruster at high velocity. ELF has been shown to operate from 10 - 100 kW, with an exhaust velocity of 15 - 40 km/s. The ELF thruster is expected to have an extremely large range of efficient power levels, high thrust density, high specific power, long lifetime, and the ability to utilize virtually any type of propellant. Thruster design and operation, novel diagnostics, and a discussion of experimental results detailing the key physical phenomena within the thruster and exhaust plume will be presented.

http://msnwllc.com/Papers/ELF_IEPC-2009-265.pdf

[Nuke]

engines like this will come in handy when we start exploring the kuiper belt in more detail. right now the way to do that is with a new horizons style flyby, which has its limits, only one or two objects at a time and no sticking around for very long. if instead you refuel at each object you explore and then pick up the propellant needed to go to the next object. mission can go on as long as the power supply holds out. it shouldn't be hard to find various ices on low gravity objects out there, you just need to set down and use your craft's waste heat to melt the ice.

[SomeGuy12]

Sounds like this is the ticket. (ELF)

Pretty much meets my checklist : high power (I would assume you would upgrade those coils to superconducting wire on a flyable version), long burn life (no or minimal electrode erosion), any propellant you want, high ISP...this is how to do it.

The lorentz force thrusters that use electrodes are currently one of the most powerful and efficient thrusters available, but they suffer greatly from erosion. This seems to fix it.

And, I guess using solids would be annoying - but if you could use water as propellant, that would be great. You would electrolyze the water to hydrogen and oxygen and store it in gas bottles. Then you'd configure the ELF for one or the other, use the pure hydrogen or pure oxygen, then switch to the other once the temporary fuel feed tank is empty.

Edited August 9, 2015 by SomeGuy12

[Streetwind]

The ELF thruster is a really neat development, not only due to the propellant thing but also due to the electric efficiency - it converts a larger part of the input energy into thrust than other electric engine types. It's a bit of a bummer that it still feels so far out. After all, most electric engines take around 25 years from first inception to flying in space, and the ELF thruster was first conceived in 2008-2009.

[SomeGuy12]

Streetwind, it's not a big deal. For electric propulsion to work, you need to fly either nuclear powered probes or probes with very large solar panels. It also has to be a mission where the huge delta-V is actually useful, such as an outer planets probe that does more than a flyby, but instead enters orbit around the planet in question.

To pull that off, you'd need the probe to be nuclear powered with a real reactor onboard, not a mere RTG. You need some legitimate high power to weight ratio. That means a compact, lightweight reactor that is still safe to launch (maybe you'd dump your fissionable material's packaging and shield after launch so you don't have to carry the weight on the journey). It means a lightweight heat engine with high power to weight ratio. Probably means a liquid tin droplet radiator, or a high performance high temperature coolant radiator.

Compared to solving all these technical problems, the electric thruster is the easy part. Some of the previous electric thrusters would have been fine - you just need an appropriate power source.

[Streetwind]

Yeah, there's no question in that regard.

But here's the thing - if developing the thruster is easy and developing a power source is hard, then performing the easy job of developing a thruster that does more with existing power sources is at least a useful step forward. Especially if we're talking about solar electric propulsion, because it means you need less giant solar panels.

And keep in mind: developing thrusters and developing power sources isn't mutually exclusive. Both happens simulatenously. So we can reap the benefit of the more energy efficient thruster even while developing new power technology at the same time. Plus, the more energy efficient thruster makes developing the power solution easier, because you can get away with something that's less sophisticated and high-performing for starters.

[Kaos]

I think this approach might be especially useful when it comes to cleaning earth orbit from debris, then you can use the debris as fuel to reach the next debris. On the engineering side I could well image that it is possible, but do not know, how complicated it really is.

[SargeRho]

I remember reading about doing that actually, a small sat would essentially "jump" from piece of debris to piece of debris, throwing each piece out of orbit.

[KerikBalm]

As far as fuel-agnostic electric engines go, there are a bunch that can theoretically eat almost anything. Just nobody does it because using xenon is simply better, and when you're dropping three-digit millions on construction and launch, the price of a few hundred kgs of xenon starts to be pretty insignificant.

Then there are proposals like the VASIMR, which is supposed to run on argon (another noble gas). Argon is flat-out worse than xenon in every respect, but it's much more common and thus much cheaper. And since the production model VASIMR is expected to start at 200kW and scale up from there - which is orders of magnitude higher than any electric engine in space today - fuel consumption is going to be higher and the company thinks that it starts making sense paying attention to the price tag of the fuel.

Well, there are a couple things wrong with what you've said.

* Argon is not worse in every respect... it is lower mass. Much like LH2 in an NTR, this means you can get higher specific impulse with argon, than with Xenon

* As you increase specific impulse, the proportion of energy going to ionize the gas becomes proportionately less. Keep in mind, after ionizing the gas, if you want to produce 1kN at 4,000 Isp, you need 1/4th the energy to produce 1kN at 16,000 Isp. The ionization energy remains the same.

* If you are relying on thermal heating (which admitedly doesn't use Xenon or Argon to begin with) like an arcjet or resistojet, then you don't need to ionize it at all.

For that matter, you could pump liquid argon through a NTR and get better performance than liquified Xenon.

The VASMIR engine operates at a significanly higher Isp than current ion drives - Dawn's ion drive runs at 3100s, whereas the VX-200 is optimized for 5000s, and can be run much higher (but it won't be as energy efficient at those higher Isps, so if it ran at 10000s, it would have less than half the thrust it gets at 5000s).

If it is very high Isp you want, Argon is better than Xenon.

I imagine that could be the case for station keeping (as on a station with very large solar arrays and can produce lots of power if other energy consuers are put on hold), or a nuclear powered craft.

That you need to send less Argon up there than Xenon, is probably more important than Xenon being cheaper than Argon.

Its $1,200 per kilgram.

Using a russian rocket to get a kg of stuff to orbit costs about $4,300

A kg of argon costs $5....

So at current (or recent... what is spaceX charging?) prices, getting the propellant there for station keeping would be 27% more expensive to use Xenon... not that much.

Xenon, being about 3.3x heavier than argon, probably results in a reduction to Isp of greater than 27% (granted, this is not like a NTR, where you have a pretty hard limit to the Isp... though you still have limits of how much current you can pump through the coils, its not as bad as an NTR operating at ~3000 K).

Thus my guess is the Isp considerations play a larger role.

I know the amount is miniscule, but it would seem kind of nice if they angled the thrust just enough that the xenon hit earths atmosphere, and was not "lost to space".

Interestingly... if you fired a NTR with an Isp of 800s on the ISS... since it orbits at about 8km/s, and the exhaust velocity would be about 8km/s... the exhaust would essentially fall straight down to Earth (while a retroburn would send it well past escape velocity).

Edited August 10, 2015 by KerikBalm

[Streetwind]

* Argon is not worse in every respect... it is lower mass. Much like LH2 in an NTR, this means you can get higher specific impulse with argon, than with Xenon

* As you increase specific impulse, the proportion of energy going to ionize the gas becomes proportionately less. Keep in mind, after ionizing the gas, if you want to produce 1kN at 4,000 Isp, you need 1/4th the energy to produce 1kN at 16,000 Isp. The ionization energy remains the same.

* If you are relying on thermal heating (which admitedly doesn't use Xenon or Argon to begin with) like an arcjet or resistojet, then you don't need to ionize it at all.

For that matter, you could pump liquid argon through a NTR and get better performance than liquified Xenon.

The VASMIR engine operates at a significanly higher Isp than current ion drives - Dawn's ion drive runs at 3100s, whereas the VX-200 is optimized for 5000s, and can be run much higher (but it won't be as energy efficient at those higher Isps, so if it ran at 10000s, it would have less than half the thrust it gets at 5000s).

If it is very high Isp you want, Argon is better than Xenon.

First and second point: yes, that effect exists; you need less energy to accelerate the smaller element to the same speed (or can use the same energy to accelerate it to greater speeds). However, this effect has its downsides. For starters, because it takes more energy to ionize the smaller element (because it clings harder to its electrons), that means that out of a finite energy budget, you have less energy remaining to spend on accelerating the propellant after ionization, even if you ionize the same number of atoms. Thus you get less than the full benefit out of argon being smaller.

Additionally, this comes at a reduction in thrust, because if you accelerate the same amount of argon atoms per time as you would xenon atoms, you accelerate less mass. Vastly less - throwing the argon faster will not compensate it. If on the other hand you wanted to have parity in thrust, you need to accelerate more argon atoms, which means each one is going to have less energy for itself, meaning it's going to not go as fast, meaning the Isp is not as high as you'd hope for. And of course, the extra argon must also be ionized, which again takes away energy that would otherwise go into accelerating propellant.

I'll admit that I don't know how exactly the numbers play out here, and would love to see some proper, hard math on it.

Finally (fourth point), thruster architecture has a far, far larger impact on thrust and Isp than the fuel choice. Two thrusters that are often compared are the VX-200 VASIMR and the Dual-Stage 4-Grid ion thruster, because they both run in similar power envelopes (~200 kW vs ~250 kW) yet represent drastically different engineering solutions. The former achieves 5 N thrust with 5,000s Isp with argon, the latter achieves 2.5 N thrust at 19,300s Isp with xenon. Argon is clearly the high thrust, low Isp option in this pairing; the argon engine gives you more than twice the thrust of the xenon engine for the same power investment, while the xenon engine is leaps and bounds ahead in terms of Isp. How does that fit together with the story that argon is better at producing high Isp? It clearly doesn't, which underlines that the choice between the two is largely unimportant in designing an electric engine for a specific performance target, and the differences are not nearly as big as they seem. I'm pretty sure Ad Astra could have used xenon as well and hit a very similar performance target. It's just that the VX-200, with its intended performance being what it is, has a significantly greater throughput of reaction mass per second than the DS4G, or indeed any other electric engine ever fired. You really don't want an expensive fuel for this kind of throughput, because the cost of filling up the tanks will get into regions that can no longer be as easily written off / ignored. This counts double if the vessel in question is supposed to be refueled and reused, like many VASIMR utilization ideas are.

(Your third point, while technically true, is not practically relevant to the discussion at hand.)

[KerikBalm]

Hmm, that hipep does sound pretty good, and yes, I acknoweldge that argon is going to get you much less thrust than Xenon for the same Isp, untill you start getting really really really high Isps - then the difference in thrust will be smaller.

"For starters, because it takes more energy to ionize the smaller element (because it clings harder to its electrons), that means that out of a finite energy budget,"

This is true, but You need to ionize 3.3x as much argon to have the same mass of ionized argon as Xenon. It takes more energy to ionize the argon, but this is minor relative to the mass difference. Argon's ionization energy is 15.75, xenon's is 12.13 -> nearly 1.3x more per atom, doesn't really compare much to the 3.3x number of atoms needed.

If you were to accelerate argon or Xenon atoms to .99C, you wouldn't really care about the ionization energy at that point.

Also, argon gives you more charge per mass, which should help you accelerate particles in a shorter distance, no?

It should also aid in reaching a higher maximum Isp... even if the thrust production at that maximum Isp is not very useful, a higher maximum Isp is an advantage... just not a very practical one.

"out of a finite energy budget, you have less energy remaining to spend on accelerating the propellant after ionization"

Yes, which is why Xenon is preferred.

I'm just saying, hypothetically, if you are spending the vast majority of your energy accelerating the propellant (such as the unreasonable case of .99c exhaust velocity), then argon should be better.

In general, as required Isp goes up, the relative advantage of Xenon gets smaller.

"if you accelerate the same amount of argon atoms per time as you would xenon atoms, you accelerate less mass. Vastly less - throwing the argon faster will not compensate it"

Well, it could, but you'd need to throw them 3.3x faster, for 3.3x the energy (1/2(m/3.3)*(3.3V)^2 = 3.3*1/2*MV^2

You'd get 3.3x the Isp at the same time.

Obviously, higher Isps come with proportionately higher power demands.

In the same mass flow+ exhaust velocity(Isp) scenario, Argon loses due mainly to needing to ionize more atoms.

The question of how badly it loses is simply a matter of what % of your energy is going to ionization.

This % gets arbitrarily low as your Isp gets arbitrarily high.

For a very very high Isp engine, argon should work fine.

If that Dual-Stage 4-Grid ion thruster really does get 2.5N at 19,300 Isp... then it would seem to be preferred to argon and the Vasmir.

Yes Xenon is expensive, but *any* mass in orbit is expensive.

You can take the price of anything on Earth, and add about $4,300/kg to it in orbit.

Xenon is then only 30% more expensive, if you get more than 30% more Isp at useful thrust due to ionization energy savings, it is cheaper to use Xenon.

But 19,300 Isp is getting quite high... I wonder how much thrust loss it would suffer from switching to argon, it may not be so bad.

Edited August 10, 2015 by KerikBalm

[Streetwind]

Thanks for the math, it'll take me some time and quiet to grok it and form an opinion, but I appreciate it nonetheless.

But 19,300 Isp is getting quite high... I wonder how much thrust loss it would suffer from switching to argon, it may not be so bad.

We probably won't find out anytime soon... ESA shelved the DS4G after initial technology validation testing because there is no usecase for it, and it never went anywhere near flight hardware. Nothing has been done with it in a decade or so. I saw a short mention that some university work managed to preserve performance while removing the fourth grid, creating a simplified dual-stage 3-grid thruster (normal gridded ion thrusters only have three grids and a single stage) but that may well have been a theoretical exercise as a PhD thesis.

There's a complex relationship between power density and Isp and desired mission profile that I don't personally understand fully (in fact I may even screw up this small description of it), which states that for a given power supply and a certain dV requirement and a certain payload, there exists an optimal engine Isp to fly this mission. And for typical applications in the Earth-Moon system, these optimal Isp numbers are only around 3,000-6,000 seconds. Ad Astra knows well what they're doing with the VX-200 being targeted at 5,000s, because that is what customers will actually want to buy. The DS4G is an impressive demonstration of technology, but it's impractical for actual usage in the near term.

The optimal Isp values go up for missions that require more dV (obviously), but also for improved power solutions. So as time goes by and solar panels improve and perhaps new non-solar power sources make an appearance, there will be demand for higher Isp engines. Something like a small onboard reactor would be a good argument to dig the DS4G back out - it would lend itself to high Isp's both through a better power density and through enabling high dV mission profiles into remote areas of the solar system where we traditionally have not been able to generate sufficient power for electric engines.

[SomeGuy12]

There's a complex relationship between power density and Isp and desired mission profile that I don't personally understand fully (in fact I may even screw up this small description of it), which states that for a given power supply and a certain dV requirement and a certain payload, there exists an optimal engine Isp to fly this mission. And for typical applications in the Earth-Moon system, these optimal Isp numbers are only around 3,000-6,000 seconds. Ad Astra knows well what they're doing with the VX-200 being targeted at 5,000s, because that is what customers will actually want to buy. The DS4G is an impressive demonstration of technology, but it's impractical for actual usage in the near term.

.

That complex relationship is any of these engines are sending a continuous stream of mass in a line out the back. The mass * velocity of this line = momentum change of the spacecraft. Momentum change is also known as "impulse". so, the effect of the engine you care about is just the equation impulse = mass_propellant * velocity_exhaust.

Well, the energy you need to accelerate the propellant to that velocity is just :

Energy_Required = (1/efficiency_engine) * (1/2 * mass_propellant * velocity_exhaust^2)

See the problem? Let's say you have 2 electric engines, and they are 50% efficient. One has double the ISP of the other. Since ISP is double, impulse is double, and therefore exhaust velocity is double. This means that Energy_required = (1/0.5) * 1/2 * (1/2)mass_propellant * 2*velocity_1^2. Or, energy requirements double for the same impulse.

That's all it is. A dual stage 4 grid, by this fundamental equation, means that since it has about 3 times the ISP, will require 3 times the energy for the same change in velocity.

Currently, that energy has to come from solar panels. Since you need 3 times the energy, but your panels haven't gotten any larger, you accelerate 3 times slower with a more efficient engine. As you know, the Dawn probe Ion engine is already crap for thrust, cutting it back by a factor of 3 (or even 6) isn't helping you any.

Solutions : if it's an inner planets probe, you could launch a bigger rocket from the ground, and basically pack in a much larger solar panel. If half the mass of the probe is a solar panel, it would be a much better power:weight ratio.

If it's outer planets, you need high power nuclear. None of this "SAFE-400" crap, I suspect the quoted 100 kilowatts/500 kilograms for that reactor is not an adequate power:mass ratio. You'd need something really pushing it, and minimal to no shielding to slow you down.

Anyways, if you had such a reactor, some of the older electric thruster designs like MPD would have been fine.

Edited August 10, 2015 by SomeGuy12

[Andersenman]

The ELF thruster is supposed to be able to operate on many different propellants.

Note that it says "any type of propellant", not "anything as propellant". It still has to be a viable propellant somehow, I don't think they mean to include stuff like wood chippings, empty beer cans or vacuum cleaner's dustbag contents.

[SuperFastJellyfish]

Note that it says "any type of propellant", not "anything as propellant". It still has to be a viable propellant somehow, I don't think they mean to include stuff like wood chippings, empty beer cans or vacuum cleaner's dustbag contents.

AFAIK, you can, but you have to turn it into plasma first. The energy needed to do this is a lot higher for beer cans than Xenon so it's just immensely inefficient. The ELF thruster uses the lorentz force to accelerate ions. It doesn't care what ions you use.

[SomeGuy12]

AFAIK, you can, but you have to turn it into plasma first. The energy needed to do this is a lot higher for beer cans than Xenon so it's just immensely inefficient. The ELF thruster uses the lorentz force to accelerate ions. It doesn't care what ions you use.

And it has to be pure plasma. As I mentioned upthread, you want the plasma to be of a single element so you can tune your engine to accelerate that particular element efficiently. So if you threw in beer cans, the engine would switch modes a few dozen times, optimized for each element in the can.