Processing/Collection of Water Ice on Mars

[BN]

Processing:

100 kg Water Ice (H2O)
Electrolysis → 11.11 kg Hydrogen (H) + 88.89 kg Oxygen (O2)
requires 555 kWh of energy

11.11 kg Hydrogen (H)
+ 61.1 kg CO2 → 22.22 kg CH4 + 49.99 kg H2O.
Sabatier Process → 44.44 kg Methane (CH4)


1 Starship Fuel/Oxidizer:
1175 tonne oxygen
325 tonne methane

8,200 MWh requiring electrolysis of 1,500 tonnes of H2O. Yields 125 tonnes excess Oxygen, 750 tonnes H2O.

If you had 100 sq meters of solar panels at 20% efficiency, at the Martian equator, it would take ~788 years to generate.

Average generation of 600-700 kW required for one Starship load of methane and oxygen in ~18 months.


[Updated 04.16.24 based on corrections by tbellman]


Collection:

Water ice sublimates in all conditions on Mars except for within craters above ~70 degrees latitude, at the lowest altitudes. Water ice concentrations in regolith range from 25% at ~70 degrees latitude to near ~100% at the north pole.



Collection will likely involve autonomous machines cutting and placing blocks of ice in pressurized containers for transport south to the Mars base. Water ice will be needed for drinking, breathing and fuel production. Significant quantities may also be needed to remove perchlorates from agricultural regolith.

At one time, there was an industry of ice collection and transport across the US. I suspect the same will be true of Mars one day.


Disclaimer: Not chemist nor scientist, also bad at math. Please point out all embarrassing errors. If there is a great thread on this specific topic, my apologies, I will collect my things and go over there. Thanks.

[tbellman]

100 kg Water Ice (H2O)
Electrolysis → 11.11 kg Hydrogen (H) + 88.89 kg Oxygen (O2)
requires 366 kWh of energy

Where did you get that energy figure?  According to Wikipedia, you need 39.4 kWh/kg of generated hydrogen at 100% efficiency, and more like 50 kWh/kg in reality.  That's 555 kWh for electrolysing 100 kg water.  (And that doesn't include energy for melting the ice; but that's minor in comparison.)

11.11 kg Hydrogen (H)
 + 33.33kg Carbon (C) [derived from 122.21 kg of atmospheric CO2)
Sabatier Process → 44.44 kg Methane (CH4)
requires ~200kWh of energy (thermally complex, so wild guess)

First of all, the Sabatier reaction is not between hydrogen and carbon, but between hydrogen and carbon dioxide (CO₂), and it does not produce just methane, but methane and water.  Half of the hydrogen goes into the water.  The real reaction is 11.11 kg H₂ + 61.1 kg CO₂ → 22.22 kg CH₄ + 49.99 kg H₂O.

Second, the Sabatier reaction is exothermic.  You need to heat the inputs, but the actual reaction produces heat.

The overall process of electrolysis plus sabatier, is:
4.5 kg H₂O + 26.5 kWh electricity + 2.75 kg CO₂  →
→  0.5 kg H₂ + 4 kg O₂ + 2.75 kg CO₂  →
→  1 kg CH₄ + 4 kg O₂ + 2.25 kg H₂O

1 Starship Fuel/Oxidizer:
750 t of Methane (min 1,500,000 kg of water ice)
2,650 t of Oxygen (min 2,704,500 kg of water ice)

Those propellant figures are for the entire stack SuperHeavy + Starship.  Only the ship part will go to Mars and need refilling of propellants.

On the other hand, the tank size of Starship seems to have increased from 1200t to 1500t, which, at a 3.6:1 ratio, would be about 1175 tonne oxygen and 325 tonne methane.

And then, you seem to be under the misunderstanding that "t" means "US short ton".  It doesn't.  "t" is the metric tonne, 1000 kg.  And yes, SpaceX uses metric units.

9,899 MWh required for the Oxidizer production via electrolysis.

I'm not sure exactly how you calculated that figure.  Using the real figures (1500 t propellant load, 25 kWh/kg of produced methane for the electrolysis), I get about 8200 MWh for a full tank load of propellant, and you need to electrolyse almost 1500 tonnes of water for that.  You wold also get an excess of about 125 tonnes of oxygen.

(Note that the water you get from the Sabatier reaction is fed back to the electrolysis stage, so you only need to harvest half that amount of ice.)

If you had 100 sq meters of solar panels at 20% efficiency, at the Martian equator, it would take ~788 years to generate.

By my calculations, you will need an average of 600-700 kW of electricity to produce a full tankload of methane and oxygen in 18 months.  In practice,you need a name-plate power of maybe ten times that, to compensate for cosine-losses, nights, duststorms; and to deal with the fact that the first half of the first time you won't be operating at full efficiency.  You will definitely need tens of thousands square meters of solar panels, yes.  That is well known.

[BN]

100 kg Water Ice (H2O)
Electrolysis → 11.11 kg Hydrogen (H) + 88.89 kg Oxygen (O2)
requires 366 kWh of energy

Where did you get that energy figure?  According to Wikipedia, you need 39.4 kWh/kg of generated hydrogen at 100% efficiency, and more like 50 kWh/kg in reality.  That's 555 kWh for electrolysing 100 kg water.  (And that doesn't include energy for melting the ice; but that's minor in comparison.)

I posted this thread at 4am last night and honestly don't remember, but your estimate seems reasonable.

11.11 kg Hydrogen (H)
 + 33.33kg Carbon (C) [derived from 122.21 kg of atmospheric CO2)
Sabatier Process → 44.44 kg Methane (CH4)
requires ~200kWh of energy (thermally complex, so wild guess)

First of all, the Sabatier reaction is not between hydrogen and carbon, but between hydrogen and carbon dioxide (CO₂), and it does not produce just methane, but methane and water.  Half of the hydrogen goes into the water.  The real reaction is 11.11 kg H₂ + 61.1 kg CO₂ → 22.22 kg CH₄ + 49.99 kg H₂O.

Second, the Sabatier reaction is exothermic.  You need to heat the inputs, but the actual reaction produces heat.

The overall process of electrolysis plus sabatier, is:
4.5 kg H₂O + 26.5 kWh electricity + 2.75 kg CO₂  →
→  0.5 kg H₂ + 4 kg O₂ + 2.75 kg CO₂  →
→  1 kg CH₄ + 4 kg O₂ + 2.25 kg H₂O

My mistake, not sure why I used C here.

You're right, the reaction is exothermic, but this is happening in a very cold environment, particularly overnight, so I expect you may still need to add energy to maintain the temperature required. I suppose this would also depend on how well the process is insulated, but I am no thermal engineer.

1 Starship Fuel/Oxidizer:
750 t of Methane (min 1,500,000 kg of water ice)
2,650 t of Oxygen (min 2,704,500 kg of water ice)

Those propellant figures are for the entire stack SuperHeavy + Starship.  Only the ship part will go to Mars and need refilling of propellants.

On the other hand, the tank size of Starship seems to have increased from 1200t to 1500t, which, at a 3.6:1 ratio, would be about 1175 tonne oxygen and 325 tonne methane.

And then, you seem to be under the misunderstanding that "t" means "US short ton".  It doesn't.  "t" is the metric tonne, 1000 kg.  And yes, SpaceX uses metric units.

Good point and thank you for the units correction. Also, I suppose we are also not accounting for any boil off losses over time.

9,899 MWh required for the Oxidizer production via electrolysis.

I'm not sure exactly how you calculated that figure.  Using the real figures (1500 t propellant load, 25 kWh/kg of produced methane for the electrolysis), I get about 8200 MWh for a full tank load of propellant, and you need to electrolyse almost 1500 tonnes of water for that.  You wold also get an excess of about 125 tonnes of oxygen.

(Note that the water you get from the Sabatier reaction is fed back to the electrolysis stage, so you only need to harvest half that amount of ice.)

Agree. Also, good point about the water reuse, that is significant.

If you had 100 sq meters of solar panels at 20% efficiency, at the Martian equator, it would take ~788 years to generate.

By my calculations, you will need an average of 600-700 kW of electricity to produce a full tankload of methane and oxygen in 18 months.  In practice,you need a name-plate power of maybe ten times that, to compensate for cosine-losses, nights, duststorms; and to deal with the fact that the first half of the first time you won't be operating at full efficiency.  You will definitely need tens of thousands square meters of solar panels, yes.  That is well known.

Dust accumulation on that kind of surface area would be a chore over time. Might require some kind of solar panel roomba to keep up.

Appreciate your input and I think I agree on all points. I will modify the OP with revised figures as they are refined in this thread. Hoping we can form a good ballpark reference for water ice resource utilization here.

[Twark_Main]

The nice thing is, on Mars insulation is easy. You just put MLI in a bag and pull a very slight slight vacuum, and you can have R-200 per inch. 

For solar panels the "roomba" is a grid of 3-4 sets of wires, which are alternately shorted to high voltage in a "chasing lights" sequence. This electrostatically sweeps dust off the panels.

https://spectrum.ieee.org/tech-from-mars-selfcleaning-solar-panels

[tbellman]

The nice thing is, on Mars insulation is easy. You just put MLI in a bag and pull a very slight slight vacuum, and you can have R-200 per inch. 

Unfortunately, that's a problem for the Sabatier reactor, not a feature.  It typically needs cooling in order to not overheat, not isolation to keep it warm...

The heat produced by the Sabatier reaction is nice, in that you can use that to heat the inputs (the hydrogen and the carbon dioxide) to a suitable temperature.  But that heat tends to be more than you need for heating the inputs, so you need to cool away the excess.

And then you need to chill the output, in order to separate the water from the methane.

And then you need to further chill the methane to make it liquid.  (And likewise the oxygen from the electrolysation step.)

(The thin and cold atmosphere is helpful in then keeping the methane and oxygen liquid, though, as isolating your tank farms becomes easier.)

(Here is one NASA design study about Sabatier reactors on Mars that I found with a quick web search.  PDF attached.  I have only had time to skim it, though, not read it thoroughly.)

[deadman1204]

How do you fuel a starship? That takes ALOT of people and equipment on Earth to do.
Astronaut with a ladder is not a soluation either.

[BN]

The nice thing is, on Mars insulation is easy. You just put MLI in a bag and pull a very slight slight vacuum, and you can have R-200 per inch. 

Unfortunately, that's a problem for the Sabatier reactor, not a feature.  It typically needs cooling in order to not overheat, not isolation to keep it warm...

The heat produced by the Sabatier reaction is nice, in that you can use that to heat the inputs (the hydrogen and the carbon dioxide) to a suitable temperature.  But that heat tends to be more than you need for heating the inputs, so you need to cool away the excess.

And then you need to chill the output, in order to separate the water from the methane.

And then you need to further chill the methane to make it liquid.  (And likewise the oxygen from the electrolysation step.)

(The thin and cold atmosphere is helpful in then keeping the methane and oxygen liquid, though, as isolating your tank farms becomes easier.)

(Here is one NASA design study about Sabatier reactors on Mars that I found with a quick web search.  PDF attached.  I have only had time to skim it, though, not read it thoroughly.)

So we would need a constant supply of CO2 and Hydrogen to sustain the reaction and avoid requiring heat input. In return, our heating bill would be much lower. It seems like that would make more sense than stopping and restarting the reactor and the output would just be chilled as we use it for heating the habitat or a greenhouse at night.

Electrolysis would be running constantly, as would ice collection operations and each step of the ice/gas resource collection and processing would ideally be automated in a single integrated system.

For solar panels the "roomba" is a grid of 3-4 sets of wires, which are alternately shorted to high voltage in a "chasing lights" sequence. This electrostatically sweeps dust off the panels.

https://spectrum.ieee.org/tech-from-mars-selfcleaning-solar-panels

"Mazumder said that within two minutes, the process removes about 90 percent of the dust deposited on a solar panel and requires only a small amount of the electricity generated by the panel for cleaning operations."

This was published 14 years ago. Is there any video or demonstration of this working on Earth?

How do you fuel a starship? That takes ALOT of people and equipment on Earth to do.
Astronaut with a ladder is not a soluation either.

I think a lot of that is the launch window and managing cryogenic fuel, pressure, temperatures and fill level.

Leaving Mars, methane would not ignite without an oxidizer, probably the fuel would not be cryogenic and the launch window would be pretty long. I don't think it would need to be as complex as most orbital launches from Earth and the Starship return variant design would likely be adapted to the Mars situation.

[Twark_Main]

The nice thing is, on Mars insulation is easy. You just put MLI in a bag and pull a very slight slight vacuum, and you can have R-200 per inch. 

Unfortunately, that's a problem for the Sabatier reactor, not a feature.  It typically needs cooling

Sure, but I don't see how "on Mars insulation is easy" is causing the problem.

Obviously if you don't need insulation, you don't install insulation.   

[BN]

Elon stated in a recent interview that they are probably landing at Arcadia Planitia.

So if we don't have automated subsurface ice harvesting, the base will certainly fail. There is no readily available, high purity exposed above-surface ice within range.


Does anyone know what this hardware looks like? It should probably be tested during the Mars mission launching next year.

Map of sites within AP and ice content chart

https://www.nasa.gov/wp-content/uploads/2015/11/viola_arcadiaplanitia_final_tagged.pdf?emrc=85d42a

[Vultur]

Elon stated in a recent interview that they are probably landing at Arcadia Planitia.

So if we don't have automated subsurface ice harvesting, the base will certainly fail. There is no readily available, high purity exposed above-surface ice within range.


Does anyone know what this hardware looks like? It should probably be tested during the Mars mission launching next year.

Map of sites within AP and ice content chart

https://www.nasa.gov/wp-content/uploads/2015/11/viola_arcadiaplanitia_final_tagged.pdf?emrc=85d42a

Ice harvesting will not necessarily be complete automated; i don't think return propellant gets made until the first humans get there (though the *hardware* to do so will be landed during the preceding cargo-only synod).

I don't think it's realistic for the 2026 synod to be that cargo-only synod. Even if Starship is interplanetary capable by Nov-Dec 2026 (which is fairly aggressive, as orbital refueling has to work first and that is only 18-19 months away) I doubt they'd be able to do more than test cruise to Mars and EDL at Mars.

[BN]

Ice harvesting will not necessarily be complete automated; i don't think return propellant gets made until the first humans get there (though the *hardware* to do so will be landed during the preceding cargo-only synod).

I don't think it's realistic for the 2026 synod to be that cargo-only synod. Even if Starship is interplanetary capable by Nov-Dec 2026 (which is fairly aggressive, as orbital refueling has to work first and that is only 18-19 months away) I doubt they'd be able to do more than test cruise to Mars and EDL at Mars.

what is it that humans would need to do that machines/tesla bot cannot do for propellant production?

[Vultur]

Ice harvesting will not necessarily be complete automated; i don't think return propellant gets made until the first humans get there (though the *hardware* to do so will be landed during the preceding cargo-only synod).

I don't think it's realistic for the 2026 synod to be that cargo-only synod. Even if Starship is interplanetary capable by Nov-Dec 2026 (which is fairly aggressive, as orbital refueling has to work first and that is only 18-19 months away) I doubt they'd be able to do more than test cruise to Mars and EDL at Mars.

what is it that humans would need to do that machines/tesla bot cannot do for propellant production?

There's a big difference between running machines real time, with access for maintenance, vs running machines with 20 minute light lag and no ability to do maintenance.

The actual harvesting would be done by machines but I'd expect humans to be heavily involved. Mars rovers move very slow, InSight had difficulties drilling, etc.; real time control would allow "industrial" operations.

[BN]

what kind of machine are you imagining? if we're just digging regolith and putting it in a hopper for separation, that can all be automated. for example NASA's RASSOR can operate autonomously.

if we're using an auger with a vacuum, there's no reason for human involvement. if there's a problem hopefully a tesla bot can reposition/reset/repair, etc

[Twark_Main]

Don't drag the ice to the sublimation oven.  Build the sublimation oven around the ice.

...

This is my general concept.

    1. Cover the surface with a thin membrane, burying the perimeter.

    2. Pull a vacuum (these two steps are already proven with vacuum surcharging on Earth).

    3. The ice underneath sublimates at the lower pressure.

    4. Re-deposit the (now clean) ice in a collection vessel.

    5. Recycle the heat of deposition back under the membrane, so it's used to sublimate more ice. Lots of heat in that phase change!

    6. Optionally you might expose ice by "gardening" with heavy equipment or blasting, make the membrane a solar collector (80% vs 25% efficient), or selectively insulate to reduce heat loss.


Fortunately this is very short range movement of water vapor (as mentioned above), so it does work...

[spacenut]

What about a small nuclear reactor, produced both heat and electricity 24/7?  Would it not be easier than a huge amount of solar panels?  Heck, the reactor could be left on board the first Starship without offloading, and the first Starship used as a fuel depot.  Then only water would have to be mined or extracted via robotics.  At this point NASA would have to get involved with SpaceX for nuclear.  It could be a molton salt type reactor using thorium.   

[Twark_Main]

Would [a small nuclear reactor] not be easier than a huge amount of solar panels?

If cost is a proxy for "ease," then no it is not easier. 

[Slarty1080]

What about a small nuclear reactor, produced both heat and electricity 24/7?  Would it not be easier than a huge amount of solar panels?  Heck, the reactor could be left on board the first Starship without offloading, and the first Starship used as a fuel depot.  Then only water would have to be mined or extracted via robotics.  At this point NASA would have to get involved with SpaceX for nuclear.  It could be a molton salt type reactor using thorium.

It would be expensive as it will require a lot of development, but IMO some form of fairly large scale nuclear plant will be required on Mars if we are ever to develop more than a "flags and foot prints" presence there. Power availability will throttle development of any surface infrastructure as it will be needed in vast quantities for ISRU production of most materials from oxygen and propellants to plastic, glass, steel and aluminum. It will also be needed to power the basic functioning of the base from vehicles and life support to heating and lighting.

[BN]

the power situation is fairly worked out. a ~50kw fission reactor will likely be used. we have already developed these as part of the Kilopower project, as well as Camp Century.

this is one of the best resources on this topic for crewed mars:

https://ntrs.nasa.gov/api/citations/20170002010/downloads/20170002010.pdf

Don't drag the ice to the sublimation oven.  Build the sublimation oven around the ice.

This is my general concept.

    1. Cover the surface with a thin membrane, burying the perimeter.
    2. Pull a vacuum (these two steps are already proven with vacuum surcharging on Earth).
    3. The ice underneath sublimates at the lower pressure.
    4. Re-deposit the (now clean) ice in a collection vessel.
    5. Recycle the heat of deposition back under the membrane, so it's used to sublimate more ice. Lots of heat in that phase change!
    6. Optionally you might expose ice by "gardening" with heavy equipment or blasting, make the membrane a solar collector (80% vs 25% efficient), or selectively insulate to reduce heat loss.

Fortunately this is very short range movement of water vapor (as mentioned above), so it does work...

don't need the membrane imo. why do you think that is necessary?

my concept:

phase 1
1. augur vacuum begins drilling
2. friction heat → localized sublimation → water vapor → condensation hopper

phase 2
1. shovel high purity regolith into a tank
2. transportation to separation hopper

phase 3
1. open pit mining of korolev ice blocks

[tbellman]

the power situation is fairly worked out. a ~50kw fission reactor will likely be used. we have already developed these as part of the Kilopower project, as well as Camp Century.

this is one of the best resources on this topic for crewed mars:

https://ntrs.nasa.gov/api/citations/20170002010/downloads/20170002010.pdf

The Kilopower project seems to have ended in 2018, after they tested their 1 kW_e prototype "KRUSTY".  No larger reactors, nor any non-prototype reactors, were developed as far as I know.  The latest plans from NASA for nuclear reactors is the Fission Surface Power, where they will attempt to buy reactors commercially.  So no, not "already developed".

As for the reactor at Camp Century, or more generally the Army Nuclear Power Program, none of the reactors designed and built were made for operation in Mars-like conditions (near vacuum, with no or extremely little air or water available for cooling).  It was also more than 50 years ago; any designs from then are not practically useful (without significant further development) today.

[BN]

the power situation is fairly worked out. a ~50kw fission reactor will likely be used. we have already developed these as part of the Kilopower project, as well as Camp Century.

this is one of the best resources on this topic for crewed mars:

https://ntrs.nasa.gov/api/citations/20170002010/downloads/20170002010.pdf

The Kilopower project seems to have ended in 2018, after they tested their 1 kW_e prototype "KRUSTY".  No larger reactors, nor any non-prototype reactors, were developed as far as I know.  The latest plans from NASA for nuclear reactors is the Fission Surface Power, where they will attempt to buy reactors commercially.  So no, not "already developed".

As for the reactor at Camp Century, or more generally the Army Nuclear Power Program, none of the reactors designed and built were made for operation in Mars-like conditions (near vacuum, with no or extremely little air or water available for cooling).  It was also more than 50 years ago; any designs from then are not practically useful (without significant further development) today.

Yes, FSP is the reactor. They're building it now based on what I mentioned above.