SpaceX Argon ion engine a game changer ?

[Michel Van]

Starlinks V2 minis are equipped with new Argon Hall thrusters for on orbit maneuvering

Argon Hall thruster tech specs:
- 170 mN thrust
- 2500 s specific impulse
- 50% total efficiency
- 4.2 kW power
- 2.1 kg mass
- Center mounted cathode

According some sources in Internet is this 2.5 times more powerful of best NASA ion engine.
But real game changer is use of cheaper Argon as propellant, (instead Xenon).
0.9% of Earth atmosphere and 1.6% of Mars atmosphere is argon.

Now this could be something for Starship equip with Argon Hall thrusters.
It could reduce transit time or able heavy cargo transport to Mars during unfavorable mission opportunities.

But real game changer is long flights to Jupiter or Saturn
With those Argon Hall thrusters you could reduce transit time from 5 years down to months.

[eeergo]

Starlinks V2 minis are equipped with new Argon Hall thrusters for on orbit maneuvering

Argon Hall thruster tech specs:
- 170 mN thrust
- 2500 s specific impulse
- 50% total efficiency
- 4.2 kW power
- 2.1 kg mass
- Center mounted cathode

According some sources in Internet is this 2.5 times more powerful of best NASA ion engine.

According to whom?

NASA-300M (developed 2004-05):
- 10-20 kW
- Up to 1.13 N
- 0.57-0.73 anodic efficiency
- 1709-3154 s ISP

SPT-190 (developed in late 90s):
- 5-30 kW
- Up to 1.5 N
- Up to 0.7 anodic efficiency
- Up to ~ 3000 s ISP

And these are strictly Hall-effect thrusters. If you widen the definition of "electric/ion thrusters" some of these specs can be better.

Of course, these specs refer to operation with xenon because it's just more efficient and the industry standard. Operation with argon will lower their specs, but we don't know by how much because we don't have the baseline against which to compare them (it is likely SpaceX's thrusters are not using pure Ar, but a mixture with small amounts of Kr or Xe that increases performances, as well as some voltage modulation "tricks"). Still, the rule of thumb is Ar performance is about 60-55% that of Xe, for the same thruster. You can see how that still leaves plenty of margin for the above US/Russian thrusters to be much more "powerful".

FWIW, argon has been used already in the 60s in space-bound thrusters, but was kept as the option of choice for ground testing because of its relatively lower performance. As noted in other threads, the Ukraine war, the sheer planned size of Starlink and other factors, such as the aforementioned recent experimentation with gas mixtures that improve raw argon performance, have prompted a revisit of this standard.
EDIT: Took info from "Lou"'s tweet (1630302082526769153) that some Starlink developers positively commented on. Thrust scales with the square root of the atomic weight, so argon's thrust should be about 55% that of xenon. However, from his analysis we know there is some "tricks" at work here, including possibly adjustments to the voltages in addition to the gas mixtures I mentioned above.

[Robotbeat]

SpaceX has long traded solar electric propulsion for their Mars architecture, and there were hints of SEP in the MCT days. It didn’t trade well enough to be included in the ITS presentation.

It wouldn’t likely improve transit times. Refueling is FAR more enabling for that. But if using Argon and extremely cheap solar panels and thrusters, it could make sense for moving propellant around and cargo. But you’re just trying to reduce the number of launches there, so you’re competing against the marginal cost of Starship launches… which is supposed to be extremely cheap. It’s a challenge to do better than that. Using Argon, however, at least makes that possible in principle whereas with Xenon or Krypton, that wouldn’t really be possible due to fundamentally higher price.

[Robotbeat]

Note that electric propulsion takes a roughly 2x delta-v hit due to being low thrust. There are games to play with that by only thrusting near periapsis, but this means you can’t use as high of Isp because you’re generating less total electricity (it generally doesn’t trade well  to use batteries to buffer the electricity for this, as you’d be better off just using bigger solar arrays or nuclear reactors).

So to first order, you cut the Effective Isp in half to compare to NTR or chemical. The dry mass of the SEP system is also usually much higher than the engine (and dry tank) mass for chemical rockets (but NTR rockets with the bulky hydrogen tanks have such high dry mass due in part to using hydrogen that it can be comparable to electric propulsion!), so you have to take that hit into account as well.

Still, if you can get high enough specific power thrusters, power supply, and solar arrays, it can trade well.

Starlink will be about 1GW when fully deployed on Starship. They will have the volume of manufacturing to make the cost low enough to make a cost competitive SEP tug system for moving cargo and propellant around.

If they can get the cost of the SEP system down to $10/Watt, I think it’d make sense versus $10/kg IMLEO Starship. BUT only over the long term. The investment of SEP would take like a decade to pay back versus just using super cheap Starship launches (and you’d have thruster lifetime issues as well).

I think Starlink v1.5 was more like $100/Watt. Starlink v2 could in principle get down to $10/Watt for SEP.

If Starship never gets below $100/kg IMLEO, then the case for SEP would be much stronger.

Argon enables SEP to work at extremely large scale whereas before SEP would be irrelevant for mass space settlement because Xenon is so scarce (Krypton not much better, overall). You’d have to take the performance hit.

(You can also use water, CO2, and nitrogen or even air for electric propulsion, of course, but it can require even bigger performance hits, thruster cleverness, and often shorter thruster lifetimes.)

[Robotbeat]

To push mass from LEO to near escape requires about 3.2km/s high thrust propulsion and roughly double that for low thrust, so 6.4km/s. Exhaust velocity (2500s*g) of 24km/s implies a wet:dry ratio of 1.3=e^(6.4/24). So a kilogram of mass to c3~0 requires about 300 grams of Argon. 300 grams of Argon that needs to be accelerated to 24km/s at 50% efficiency requires: .5*.3kg*(24km/s)/0.50 = 172.8MJ.

If we assume the tug could return to LEO using propulsion less aerobraking over a couple months, then the process can repeat.

If we assume it takes a year (8760h*3600s=~pi*10^7 s) to do that 172.8MJ, then that needs 5.5 Watts on average. Assume it’s in shadow about 23% of the time on average, and so it requires 7Watts of thruster. That implies 14Watts of solar panels. Thruster is 2000W/kg, so 3.5grams of thruster, if solar array is 150W/kg then that implies 93 grams of solar array plus let’s say another 3.5 grams of electronics, so 100 grams total of SEP mass. 300 grams of argon implies about 50-75 grams of tank mass, less if you cryogenically chill the argon (doesn’t have to be fully liquid, but if you made it liquid, tank mass could be just 10 grams).

So if that 1000 grams final mass, around 150 grams was parasitic SEP mass, 850grams of payload (assume he spacecraft bus is negligible). Plus the additional 300 grams of Argon.

To push 850 grams  of payload using chemical propulsion using methalox would need roughly 1400 grams of methalox propellant (the chemical tug mass would be about 100-150 grams of that), but could be done in days or weeks instead of a year.

If we assume the SEP tug is used 10 times, then our 150gram tug used 3000grams of Argon displaced about 14,000grams of LEO methalox, or a net of 11kg roughly.

If that 14Watts of input power cost $10/Watt, then it cost $140 to save 11kg (the raw costs of 3kg of Argon and 14kg of methalox are similar on the ground BTW… Argon is $0.6/kg, methalox is $0.25/kg, so the Argon is $1.80, the methalox $3.50). I’m assuming the methalox tug is basically free as it can be reused so often. So the cost to beta would beat is about $13/kg to orbit, neglecting cost of capital.

If you include cost of capital (ie the interest rate), then you probably need the IMLEO launch cost to be more like $20/kg for SEP to make sense. Or you reduce the cost to $5/W input (in which case it can compete with $10/kg IMLEO)

You can also use it longer. But longer than 10 years doesn’t help total cost unless your interest rate (cost of capital) is extremely low and you have very long lifetime components.

Anyway, yeah, Argon actually DOES make a difference in the trades here.

Note that in the above, you can possibly get a better payback if you REDUCE the Isp of the Argon thruster as it reduces the amount of energy you need to pump into the propellant. It’s a tough trade, as reducing Isp reduces energy efficiency somewhat and increases propellant (and tank dry) mass, but it could work, especially if you’re using liquid Argon. Energy of the total exhaust mass is proportional to propellant mass in linear terms but exhaust velocity in squared term.

[Robotbeat]

Just to compare, if you used Krypton ($300/kg optimistically… ignoring global production constraints) in the above system, the propellant cost alone would be ~7 times the cost of the SEP tug, and it’d never beat Starship unless Starship never gets below around $160/kg (or $24 million per launch, about the internal price SpaceX pays for a Falcon 9 launch).

Argon is so cheap that it’d be cheaper than the *ground price* for the equivalent delta-v of methalox, so it’s potentially what could enable SEP to compete.

Note that solar panels can cost as low as 20-30¢/Watt on the ground. So $5/Watt-input ($10/Watt output) for the SEP system is not out of the question at all if SpaceX can mass produce the thrusters and electronics. SpaceX at least originally used like SunPower silicon cells for Starlink, not the hyper-expensive Gallium Arsenide triple junction cells from Spectrolab which are like $50-100/Watt at the cell level.

[jimvela]

I don't know any of the specifics of the SpaceX thruster.

That said, I am working a SEP mission for deep space, and the SEP power and propulsion avionics that is under trade for that application has a wide configurability for thruster performance. 

It's possible to trade all these parameters to achieve mission objectives- allowing exactly the kind of varying of power inputs, flow rates, and SEP thruster settings.

I have doubts about argon throughput for a really long-time use out in deep space, but pushing the state of the art in this regard will have big implications for all future missions.

I makes me wonder if at some point SpaceX will target the next step over in the RocketLab model- hosting payloads on a SpaceX mission bus.

The starlink bus appears to be something special.

[Robotbeat]

For deep space robotic missions, Argon doesn’t trade well against Krypton unless your mission mass is sensitive to $300/kg. A 5 ton probe carrying 2 tons of propellant is likely going to not care about the $600,000 difference in propellant costs as it’d likely be much less than 1% of the mission cost but would require about 10% more bus power and a bit more dry mass.

Xenon might be expensive enough to pick Krypton for such missions, but I don’t think Argon would make the cut unless you’re doing something super crazy like a 100 ton monstrosity that’s mostly propellant and is relying on Starship for low launch costs (in which case, good luck negotiating with SpaceX!).

The move from Xenon to Krypton made sense if you get Falcon 9 level launch costs ($1000/kg). The move from Krypton to Argon only really makes sense if you can get Starship level launch costs ($300/kg or lower). Or if you simply need a huge amount of propellant that couldn’t be supplied on the global markets.

[DanClemmensen]

The move from Xenon to Krypton made sense if you get Falcon 9 level launch costs ($1000/kg). The move from Krypton to Argon only really makes sense if you can get Starship level launch costs ($300/kg or lower). Or if you simply need a huge amount of propellant that couldn’t be supplied on the global markets.

Launch costs are driven by the lowest bidder. What year are you designing for? Starting no later than 2027, you have "Starship level launch costs".

As a separate issue, there is a potential for an orbital Argon economy, where SpaceX maintains an Argon depot and spacecraft can refuel or be refueled in orbit.

See
  https://en.wikipedia.org/wiki/Atmosphere_of_Earth
Argon is the most common constituent of air after Nitrogen and oxygen. It's more common that CO₂. A gas plant that produces pure liquid Nitrogen and Oxygen must also produce Argon as a byproduct. For every 100 tonnes of Oxygen you get about 2 tonnes of Argon, so the plant that is supplying LN₂ and LOX for Starship launches will supply far more Argon than its payloads will ever need.

[Robotbeat]

One way to look at the trade for SEP is to consider the cost per Watt of solar in orbit versus the ground. Starship is supposed to use solar produced methalox eventually. It takes roughly 3MJ of solar electricity to make 1MJ of methalox, BUT solar Only has a capacity factor of like 20% at the DC level, but in orbit it’s like 60-100% (depending on orbit). So 1 watt in orbit is worth about 10 Watts making methalox on the ground. Plus Starship is not fully efficient. It takes about 500MJ per kg to orbit whereas the theoretical minimum is 32MJ/kg (both kinetic and potential energy). Earlier variants like ITS got 320MJ/kg, and that’s probably achievable long-term. So you have another factor of 10 to the advantage of orbit.

HOWEVER, electric propulsion is using earth-launched propellant and so it can’t operate at the naive energy-optimum Isp. It has to pump about 10 times as much energy into the exhaust as optimum as its Isp is about 10x that of the mission delta-v for pushing payloads to higher orbits. Additionally, electric thrusters (like chemical rocket engines in vacuum) are only about 50% efficient. So that’s a factor of 2. Plus the low thrust means you lose out on another factor of 2 due to lack of much Oberth effect. So you’re at a factor of 40x more energy that you’d theoretically want.

40x is still better than 100x, but not by much. The fact that low cost solar arrays and infrastructure on Earth might be able to access lower cost of capital and not have to deal with the Van Allen belts could make up most of that difference.

Still, SEP does allow offloading of about half the energy of launching a massive Mars campaign to space, and that’s a big help if you start running into atmospheric constraints on launch rate. And it’s possible those atmospheric constraints come into play before you get all the way down to $10/kg with Starship.

That Argon could also be retrieved from Mars and that you could use SEP for round trips could also potentially help lower the cost further. (Instead of 25 launches for 500 Mars settlers on 5 crewed Starship launches, you could do one Earth launch of 500 settlers who transfer to 5 Mars-launched and Mars-serviced Starships which refuel in high Earth orbit with Mars-produced propellant that was transferred by Mars-refueled SEP tugs.)

[Robotbeat]

The move from Xenon to Krypton made sense if you get Falcon 9 level launch costs ($1000/kg). The move from Krypton to Argon only really makes sense if you can get Starship level launch costs ($300/kg or lower). Or if you simply need a huge amount of propellant that couldn’t be supplied on the global markets.

Launch costs are driven by the lowest bidder. What year are you designing for? Starting no later than 2027, you have "Starship level launch costs".

As a separate issue, there is a potential for an orbital Argon economy, where SpaceX maintains an Argon depot and spacecraft can refuel or be refueled in orbit.

See
  https://en.wikipedia.org/wiki/Atmosphere_of_Earth
Argon is the most common constituent of air after Nitrogen and oxygen. It's more common that CO₂. A gas plant that produces pure liquid Nitrogen and Oxygen must also produce Argon as a byproduct. For every 100 tonnes of Oxygen you get about 2 tonnes of Argon, so the plant that is supplying LN₂ and LOX for Starship launches will supply far more Argon than its payloads will ever need.

But if SpaceX is making methane using Sabatier on Earth, that implies they’re also make oxygen at a stoichiometric ratio and it’s more efficient to liquefy and subcool methane and oxygen directly than by using an LN2 intermediary, so it’s possible that long term they wouldn’t be doing a ton of air liquefaction except on Mars (where nitrogen and CO2 are needed… on Earth you’d use a different process than liquefaction to extract CO2 from the atmosphere).

[Blackjax]

As a separate issue, there is a potential for an orbital Argon economy, where SpaceX maintains an Argon depot and spacecraft can refuel or be refueled in orbit.

See
  https://en.wikipedia.org/wiki/Atmosphere_of_Earth
Argon is the most common constituent of air after Nitrogen and oxygen. It's more common that CO₂. A gas plant that produces pure liquid Nitrogen and Oxygen must also produce Argon as a byproduct. For every 100 tonnes of Oxygen you get about 2 tonnes of Argon, so the plant that is supplying LN₂ and LOX for Starship launches will supply far more Argon than its payloads will ever need.

Seems unlikely to have played a role in the decision to develop this tech  but it's worth noting that if your technical infrastructure is based around using Argon it's also applicable to the Martian atmosphere as well.

[Robotbeat]

Speaking of, I’m pretty sure the 265ton per day Air Liquide plant on Merritt Island is too small for the 100 Starship launches per year, which would need at least 1100 tons per day.

There’s another air liquefaction plant of about that size in Orlando. Plus a Praxair (now Linde) one a bit closer to KSC. BUT I think all of them combined would struggle to supply that many Starships, which require 4000tonnes of LOx per launch. Heck, even Falcon 9 requires a lot, like 400t per launch, and Falcon Heavy about 1000t. So 100 Falcon launches per year could require like 40,000t per year, or over 100tpd (tons per day) alone, even if 20 of them were from California.

SpaceX must be planning some source for a lot more LOx and LN2 to support Starship launches.

[Blackjax]

Starlinks V2 minis are equipped with new Argon Hall thrusters for on orbit maneuvering

Argon Hall thruster tech specs:
- 170 mN thrust
- 2500 s specific impulse
- 50% total efficiency
- 4.2 kW power
- 2.1 kg mass
- Center mounted cathode

According some sources in Internet is this 2.5 times more powerful of best NASA ion engine.

According to whom?

NASA-300M (developed 2004-05):
- 10-20 kW
- Up to 1.13 N
- 0.57-0.73 anodic efficiency
- 1709-3154 s ISP

SPT-190 (developed in late 90s):
- 5-30 kW
- Up to 1.5 N
- Up to 0.7 anodic efficiency
- Up to ~ 3000 s ISP

And these are strictly Hall-effect thrusters. If you widen the definition of "electric/ion thrusters" some of these specs can be better.

Of course, these specs refer to operation with xenon because it's just more efficient and the industry standard. Operation with argon will lower their specs, but we don't know by how much because we don't have the baseline against which to compare them (it is likely SpaceX's thrusters are not using pure Ar, but a mixture with small amounts of Kr or Xe that increases performances). FWIW, argon has been used already in the 60s in space-bound thrusters, but was kept as the option of choice for ground testing because of its relatively lower performance. As noted in other threads, the Ukraine war, the sheer planned size of Starlink and other factors, such as the aforementioned recent experimentation with gas mixtures that improve raw argon performance, have prompted a revisit of this standard.

This raises a good point, in order to really consider the practical impact of the new engine it'd be helpful to have a roundup of the current state of the art available out there.  Other than this wikipedia entry

https://en.wikipedia.org/wiki/Ion_thruster?wprov=sfla1

Is anyone aware of any articles/studies which roundup/review what the competition which could actually be intetegrated into an upcoming spacecraft might be?

[Robotbeat]

I pointed it out in the other thread. The Busek Hall thrusters are most comparable and are integrated into the OneWeb satellites.

[DanClemmensen]

As a separate issue, there is a potential for an orbital Argon economy, where SpaceX maintains an Argon depot and spacecraft can refuel or be refueled in orbit.

See
  https://en.wikipedia.org/wiki/Atmosphere_of_Earth
Argon is the most common constituent of air after Nitrogen and oxygen. It's more common that CO₂. A gas plant that produces pure liquid Nitrogen and Oxygen must also produce Argon as a byproduct. For every 100 tonnes of Oxygen you get about 2 tonnes of Argon, so the plant that is supplying LN₂ and LOX for Starship launches will supply far more Argon than its payloads will ever need.

Seems unlikely to have played a role in the decision to develop this tech  but it's worth noting that if your technical infrastructure is based around using Argon it's also applicable to the Martian atmosphere as well.

You are correct: the decision to use Argon is likely based on its cost (cheap) and the cost/kg to orbit (projected to become dramatically lower). But when you are projecting the cost of a commodity you need to consider whether or not  your new demand will drive up the cost: Lithium is a recent example. I was attempting to show that the cost of Argon will not rise due to scarcity, because its supply will more or less automatically rise with the increase in demand for LOX.

[virtuallynathan]

https://twitter.com/longmier/status/1630626524318797825?s=20

[wannamoonbase]

I don't know about this specific engine, but ion propulsion should have a huge future for moving around a lot of on orbit mass.

The SpaceX propellant depot could use argon engines to maintain it's orbit, or even spend months sending fuel and cargo to the moon. 

A long slow trip from the Earth to the Moon could help with refueling and reusing the HLS one day. 

That's what I want, massively more cost effective access to the lunar surface.

[tbellman]

I don't know about this specific engine, but ion propulsion should have a huge future for moving around a lot of on orbit mass.

The SpaceX propellant depot could use argon engines to maintain it's orbit, or even spend months sending fuel and cargo to the moon. 

A long slow trip from the Earth to the Moon could help with refueling and reusing the HLS one day.

When you say "months", you really mean "years", or possibly even "decades", right?

The normal delta-v from LEO to NRHO is about 3.6 km/s.  If we take one hundred of these thrusters, so we get a total thrust of 17 newton and using 420 kW electricity, and apply that to a spacecraft massing 250 tonnes, it will take around 615 days to achive that Δv.  But unfortunately, as we are using very low thrust, the amount of Δv needed increases to around 7 km/s, which will take three years and three months with 17 N thrust.  (And here I have ignored that at the start of the trip, the spacecraft will be 72 tonnes heavier, due to the argon propellant.)

If you fill the ship with more methalox to be delivered to NRHO, the time increases.  At 1800 tonnes total mass, the travel time is a bit over 23 years -- still ignoring the 500 tonnes of argon needed.

To get down to "months" for the trip from LEO to NRHO for a Starship propellant tanker or depot, you would need around a thousand of these ion thrusters.

Using ion thrusters for that purpose would certainly be "game-changing"; but not in a way that most of us would prefer...

[Robotbeat]

I don't know about this specific engine, but ion propulsion should have a huge future for moving around a lot of on orbit mass.

The SpaceX propellant depot could use argon engines to maintain it's orbit, or even spend months sending fuel and cargo to the moon. 

A long slow trip from the Earth to the Moon could help with refueling and reusing the HLS one day. 

That's what I want, massively more cost effective access to the lunar surface.

If I play the numbers just right, it’s possible to get it cost competitive with $100/kg, and maybe even $10/kg if they can tolerate an effective annual payback rate of like 10% and if they can get the cost of the spacecraft down to almost commodity prices (like, price of residential solar roofs for the solar panels, $2/Watt for the thruster, etc). But it’s not a huge improvement compared to the improvement from rapid, full reuse.

If the use case is pushing cargo or propellant from LEO to near-escape velocity, then Argon thrusters could reduce the mass required by about a factor of 2 compared to methalox refueling. But the cost isn’t half, because of course a big SEP tug would cost money.

Still, it might help things, also acts as insurance in case Starship plateaus at $100/kg instead of the hoped-for $10/kg.