The Space Reviewin association with SpaceNews

Terraformed Mars
Terraformed Mars being greened with a nitrogen-oxygen atmosphere. (credit: Kevin Gill)

Oxygen for Mars

Bookmark and Share

There is a lot of attention in our community on creating a backup location for humanity and, along with pressurized in-space settlements, Mars is one of the best locations for that. But along with the human race and its civilization, we should also include the important requirement that we need a backup for life itself. Right now the only place known where life exists, and can survive and grow, is Earth, just one tiny planet. Our lives are enabled and enriched by the incredible variety of animals and plants that live on Earth, on land and in the seas. We would be immensely impoverished if all we had kept alive in space were the most critical plant species that we grow for food. So we need to think more about how to keep a whole biosphere alive.

In a 2015 article, Don Barker made one of the best statements in support of Mars settlement I have seen, and he mentioned a “backup for life on Earth”:

“The settlement of Mars is probably the most viable endeavor that would create a backup for life on Earth, in effect a global-mitosis. Mars is the only destination whose environment and accessible natural resources efficiently enable permanent and sustainable human habitation off Earth.” — “The Mars Imperative”, Barker, Donald C., Volume 107, Acta Astronautica 3-2015.

What is really needed is to create a habitat that is very large, one that has the physical land surface area equivalent to the outdoors on Earth instead of just pressurized habitats. To match some level of open space for significant numbers (breeding populations) of even a few thousands species would require dozens or hundreds of very large, dedicated artificial habitats in space. I hope such habitats can be built, and such habitats can serve as interim refuges, but having a fully terraformed planet where the animals and plants can live in the vastly larger space outside of habitats is very important. That requires terraforming Mars, the only planet in our solar system where terraforming will be practical without “magic physics.”

For practical terraforming efforts of almost any kind, there are two basic requirements: large scale access to space and the destinations in it, and some form of fusion energy for power and propulsion.

Some people criticize discussions of terraforming as “grandiose” or a “display of hubris” and being at an “industrial-scale,” faintly echoing the “Small is Beautiful” movement of 40 years ago. If you are trying to modify an entire planet’s atmosphere, the scale is by necessity “grandiose.” These anti-tech attitudes remind me of the desperate attempts by radicals to stop the California Condor captive breeding program. It seemed as if the radicals would rather have had the condors go extinct than allow a practical human effort to save a species to succeed. That program has resulted in a huge increase from 22 condors in 1982 to 540 in 2023, an almost 25-fold increase in just 41 years.

The same kind of values are being pitted against each other here, but on a vastly larger scale: the preservation of a major part of the vast diversity of life itself versus keeping the status quo for dead rocks on what currently is probably a lifeless and desolate planet. Even if there were living microbes deep under the surface of Mars, the value of millions of multicellular species would surely outweigh the value of microbes.

I have already covered the issue of how to provide the bulk of air pressure needed to terraform Mars (see “Rethinking the Mars terraforming debate”, The Space Review August 20, 2018). That article describes how huge amounts of solid nitrogen can be moved from Pluto or other locations and dumped into the Martian atmosphere with no cratering. This would allow rain to fall and liquid water to flow on Mars. Several people have suggested Venus as a closer source of nitrogen, but it is at the bottom of a deep gravity well, and would have to be skimmed off the top of the atmosphere, separated from the carbon dioxide and liquefied for shipment to Mars with a significant delta-V value. Pluto has a shallow gravity well and the nitrogen is already in solid form, and thus much easier to mine and ship in bulk.

The two things we need for full terraforming of Mars are nitrogen pressure and some oxygen pressure. Only anaerobic bacteria could live on Mars without oxygen. Therefore, Mars also needs oxygen.

For practical terraforming efforts of almost any kind, there are two basic requirements: large scale access to space and the destinations in it, and some form of fusion energy for power and propulsion. We should soon have the first, and things are looking promising to have the second. Without fusion, both terraforming and interstellar travel would be almost impossible. Since fusion occurs in nature inside every star, it is not magic physics. The huge amounts of fusion energy needed to transport the nitrogen to Mars and create the oxygen on Mars are comparable.

Wanted: 880 trillion tons of oxygen gas to support plant and animal Earth life on Mars. This amount would provide an average of three pounds per square inch (psi) of partial oxygen pressure for the whole planet. This is the normal sea level amount of oxygen we breathe, since most of the rest of the 14.7 psi we breathe on Earth is nitrogen. About the highest people can function is at about 5,200 meters (17,000 feet), where the oxygen pressure is about 40% of sea level, or about 1.2 psi, and those people are very highly adapted. Some people get altitude sickness at only 2,700 meters (9,000 feet), where the oxygen pressure is 70% of sea level or 2.1 psi partial pressure.

How do we know how much oxygen it would take to oxygenate a whole planet? Mars has 144.8 million square kilometers of surface, and each square kilometer covers one million square meters. The nitrogen pressure must always be several times higher than the oxygen pressure to prevent a runaway oxygen fire.

There are at least three ways to measure the amount of oxygen needed on Mars: pressure per square inch, pressure per square meter, or mass per square meter (which means all of the oxygen in the “air column” directly above each square meter, measured from the surface to the top of the atmosphere.) To provide 3 psi of oxygen air pressure on Mars, with its lower gravity, takes about 2.1 (Mars) tons of oxygen pressure per square meter or 2.1 million tons per square kilometer. But since Mars gravity is about 0.376 of a G, you would need 2.66 times more oxygen to get the needed pressure, or about 5.59 actual tons above each square meter of surface. The 2.1 tons per square meter is also the amount of oxygen pressure you would get on Mars. If Mars had the same gravity as Earth, you would need only 304 trillion tons of oxygen. So if each square kilometer needs 5.59 million tons of oxygen, Mars needs 5.59 million tons times 144.8 million square kilometers, which equals about 810 trillion tons of oxygen.

In nature, oxygen is a strongly reactive element like fluorine and “wants” to combine with other elements to create compounds like water. Thus there are no oxygen mines in nature, unless you want to rob Earth’s biologically-created atmosphere of a significant part of its oxygen. So the oxygen must be chemically created from an oxygen compound like water.

In fact, Mars has many thousands of cubic kilometers of water ice, each containing one billion tons or one billion cubic meters of water ice. If we have on Mars at least one million cubic kilometers of water ice, a volume of ice regolith or glacier 1,000 by 1,000 by 1 kilometer deep, that amount would contain 1,000 trillion tons or 1 quadrillion tons of ice. If this is pure ice, it would be under an area of one million square kilometers. (Mars actually has about 145 surface areas this large.) The 1,000 trillion tons of ice can be converted to 880 trillion tons of oxygen and 120 trillion tons of hydrogen via hydrolysis, exceeding the desired amount of oxygen by about 10%.

If there turns out to be less than the needed amount of water ice on Mars, massive amounts can also be imported from the abundant “iceteroids” (asteroids mostly made of ice.) Once we have fusion power, we should be able to provide energy in the extremely large amounts needed to provide Mars with the needed oxygen. Along with the imported nitrogen, this will allow full terraforming of Mars.

Water electrolysis (using the alkaline electrolysis method) has an effective electrical efficiency of 70–80% and produces one kilogram of hydrogen and about nine kilograms of oxygen from ten kilograms of water ice. This requires 50,000–55,000 watt-hours (180–200 megajoules) of electricity. Using the same method, it takes five megawatt-hours to electrolyze one metric ton of water.

Mars has many thousands of cubic kilometers of water ice, each containing one billion tons or one billion cubic meters of water ice.

To electrolyze that one metric ton of water takes 6,300,000 watt-hours using the PEM (Proton Exchange Membrane) method as used on the International Space Station (based on information from the Hydrogen Based Energy Conversion Handbook). This is a difference of almost 20%. (Note that I am using the higher PEM value here.) Since the ratio, by mass, of hydrogen to oxygen is 2:16, then oxygen is 16/18 = 8/9ths of a given unit mass of water. Therefore ,there are about 888 kilograms of oxygen and about 112 kilogram of hydrogen in 1 metric ton of water.

The Mars water ice input requirement to generate 3 psi of oxygen is 1,000 trillion tons or 1 quadrillion metric tons. The required ice volume would be about 1,000 by 1,000 by 1 kilometer deep. The hydrolysis plant is assumed to have an efficiency of about 71%. Production numbers are based on the use of about 6,300 kilowatt-hours or 6.3 megawatt-hours to electrolyze one metric ton of water as above. Condensing the vast bulk of the resulting gas into cryogenic liquids is not needed, but I will assume that the plant would use 7.0 megawatt-hours for each ton of water to allow about an 11% margin.

Here is a scale-up table showing energy use and mass for hydrolysis of different amounts of water ice. Each line or row has 1,000 times (three orders of magnitude) more of mass and energy use than the last. Most people are not even familiar with the names of amounts of energy this large.

Energy use (amounts) Water ice mass Oxygen Mass Hydrogen Mass
7 watt-hours 1 gram 0.88 grams 0.12 g
7 kilowatt-hours 1 kilogram 0.88 kg 0.12 kg
7 megawatt-hours 1 metric ton 0.88 tons 0.12 ton
7 gigawatt-hours 1 thousand tons 880 tons 120 tons
7 terawatt-hours 1 million tons 880,000 tons 120,000 tons
7 petawatt-hours 1 billion tons 880 million tons 120 million tons
7 exawatt-hours 1 trillion tons 880 billion tons 120 billion tons

The last row in italics shows the quantities we actually need to deal with.

The extra 8–10% (80 trillion tons) of oxygen would possibly be absorbed by the surface rocks and regolith after billions of years of exposure to a near vacuum had reduced the oxygen pressure and molecular rock surface saturation levels to effectively zero.

So, over what time scale is it reasonable to expend the seven zettawatt-hours of energy? To split this up into manageable sized components, we could start arbitrarily with 7,000 plants, each one of which would expend “‘only” 1 exawatt-hour during the operational period. This large set of installations could be named the Edgar Rice Burroughs Oxygen Production System or ERB-OPS, since he described a huge “atmosphere plant” existing on Mars in his famous Barsoom stories written about a century ago. This plan assumes we can eventually build very large fusion power plants.

So, over what time scale is it reasonable for each plant to each expend its 1 exawatt-hour of energy? The larger portion of the full, breathable air pressure on Mars should be able to be recreated in about 150–200 years via the simultaneous import of nitrogen from the outer solar system. If we set the operational E.R. Burroughs oxygen plant project period at 200 years, each plant would need to expend five petawatt-hours per Earth year. (one exawatt-hours / 200 = five petawatt-hours). The exact time period is not terribly critical and would be based on future decisions about the cost and very large scale of the terraforming operation.

Since each Earth year has 8,766 hours, the energy at each plant needs to be expended at a rate of 5,000,000,000,000,000 / 8766 = ~570,400,000,000 watt-hours or 570.4 gigawatt-hours per hour, or a power level or rate of 570.4 gigawatts. During full-scale operation of the Burroughs installation, all 7,000 plants, running at 570.4 gigawatts per hour, would be expending just about 4,000 terawatts or four petawatts. (Due to the high range of magnitudes in the calculation, the single plant rate should be between 570.4 and 571.5 gigawatts).

If all humanity is currently using about 20 terawatts or 20,000 gigawatts, the level of power from one Burroughs plant would be at the same level as about 1/35th of all current human energy production (570 / 20,000 = 0.0285), while the total Burroughs system power level would about 200 times larger than all of current human energy production and use (4,000 /20 = 200). For another comparison, the Earth receives, as a constant rate, 174,000 terawatts of solar energy on is sunward side. About 122,000 terawatts of this (122 petawatts per hour) reaches the surface and lower atmosphere. For every 24 hours, the Earth thus receives 2,923,000 terawatt-hours from the Sun. So the Burroughs system would be generating and using power equivalent to about 3% of the sunlight the entire sunlit side of the Earth gets.

In the table below, all measurements are in either terawatts (1,000 gigawatts) or terawatt-hours. For example, North America uses an electrical power rate of several terawatts. So each plant would be producing power at a rate about 1/10th of that or about 500 times more than a typical one-gigawatt power plant. Each plant of course can exist as a set of ten, 50 gigawatt plants close together.

System: Sun to Earth Burroughs system 1 Burroughs plant Mankind 2023
Power Level 122,000 terawatts 4,000 terawatts 0.570 terawatts ~20 terawatts
Amount/Day 2,923,000 TWH 96,000 TWH 13,680 TWH 480 TWH
Amount/Year 1,069,000,000 TWH 35,634,000 TWH 5000 TWH 175,320 TWH
Amt /200 Yrs 213,890,000,000 TWH 7,012,800,000 TWH 1,000,000 TWH 35,064,000 TWH

You would think this might overheat the whole planet, but an area on Mars only gets half of the sunlight Earth gets, so even if the heat was distributed over the whole planet evenly, Mars would not cook. To even be as warm as the Earth is without added greenhouse gases, it would need to get about 26,000 terawatts of sunlight on its quarter-Earth-size sunward face, while it currently gets only about 18,000 terawatts. The areas being mined for water ice and supporting the plants would amount to about 1/145 of the surface, or one million square kilometers out of 144.8 million square kilometers. In practice, the equipment to build the plants would build them serially, moving from site to site, so that they would all come on-line over a period of decades.

Such high levels of energy use could only be practically supplied by fusion power. We also assume that these fusion plants will be able to be scaled up from a single city size (one to five gigawatts) to produce the most efficient size of fusion reactor, so that the power from one or several reactors can be distributed to the multiple hydrolysis plants surrounding each reactor using the least amount of power cable and associated mass. It is hoped that a form of aneutronic (no neutrons) fusion, such as boron-hydrogen fusion, will be available when the plants are designed.

The time period for reaching a breathable oxygen/nitrogen atmosphere on Mars can be less than it took for some Gothic cathedrals to be built.

Efficient design of the fusion plants and hydrolysis plants can also reduce the amount of heat released. Early work is currently ongoing to try to generate power directly from the fusion process itself, so that there would not have to be a standard heat engine with turbine generating power at each plant. In addition, the electrolysis process used here as the main example may be able to be improved massively, not simply by improving liquid hydrolysis, but possibly using super-critical temperature electrolysis, steam electrolysis, catalytic electrolysis, or other advanced methods. So the total power requirement may be much less than the values given here for the scale-up of a standard system. There are also different methods used for liquid hydrolysis. Some use proton exchange membranes and others use added electrolytes such as sodium and sulfur to allow the electric current to move through the water from one electrode to another. Mars has plenty of both sodium and sulfur. Unfortunately, there are few sources that can give us a clue on how much the future energy savings might be. In a future of abundant energy, the design effort would not be to just reduce energy use, but to reduce the amount of materials used and potentially to reduce the scale of the effort.

Assume that the ice layer areas are being mined to an average of about one kilometer deep at plant locations, with one billion metric tons of ice under each square kilometer. If the ice layer is not pure ice, the final excavation layer would be deeper or larger. The operational mode would probably be similar to that for a surface mine for copper or iron, looking like a concentric set of circles, and preventing too steep a slope in the work area during mining operations. With shallower ice deposits or those with layers of gravel, the total area would be wider. The fusion and hydrolysis plants would be located on nearby areas that are not mined, preferably those with little or no ice under them. Crew quarters would be almost entirely underground for thermal and radiation protection. Crews would have access to the plants via pressurized tunnels and control rooms. The oxygen produced would be released directly into the atmosphere, and would possibly be able to carry some or most of any waste heat quickly up and away from the industrial sites. Significant amounts of the hydrogen from plants closer to settlements would go through pipelines to be used as rocket propellant or as feedstock for making methane.

Of the total of one million cubic kilometers of water ice, each of the 7,000 fusion and hydrolysis plants would process a total of about 143 cubic kilometers of ice, or a total of 0.715 cubic kilometer (715 million tons of ice) per year. Such a mining system would cover about 1/145th of the total area of Mars (which is 144.8 million square kilometers), probably spread around the sub-polar areas in a circle so we do not disturb the layered ice deposits there right away. At the end of the operational electrolysis period, the hydrolysis plants, covering a small fraction of each process area, would be converted to other uses and the land area freed up. If the ancient Borealic Ocean is replaced eventually, the excavated areas would have deeper water and there should also be some deep basins around the South Pole.

If both nitrogen transfer and oxygen generation start to take place, we need to understand how life would slowly start to take hold on the surface of Mars. Currently, with no liquid water possible on the surface due to the very low pressure, (effectively a physiological vacuum), with constant cosmic radiation and ultraviolet radiation and with virtually all the water on Mars frozen, there are probably no active life processes occurring on or near the surface.

If we assume that Mars has been warmed enough by terraforming efforts to sublime the remaining 25 trillion tons of dry ice at the south pole of Mars, doubling the amount in the atmosphere to 50 trillion tons, this alone would raise the boiling point of water to about 7 degrees Celsius., allowing liquid water in small areas at lower elevations. It would also raise the carbon dioxide pressure level to about 1.2 millibars average from the current 0.7 millibars. Current carbon dioxide levels on Earth are about half a millibar or 0.42 parts per thousand. For breathing, carbon dioxide should be no more than about one part per thousand or one millibar in an atmosphere, but the 20% extra carbon dioxide pressure should cause no problems.

By the time we have added ten times this amount of nitrogen or 500 trillion tons, water could be a liquid at many elevations, allowing clouds to form and rain would probably fall. With some rain, the dust would slowly be washed out of the air, and streams and rivers would start flowing. The sky would gradually turn from creamy-tan to a very dark blue. The rain would start to wash the perchlorates out of the regolith and eventually into the Borealic Ocean. Since the amounts of perchlorates in the regolith could be as high as 1%, this could take a while. Until the bacteria species that can digest the perchlorates (such as some in the genera Dechloromonas and Azospira), are introduced into each stream or river, cyanobacteria would not be able to grow there. Spores of bacteria may be able to be widely introduced by the wind, so that the headwaters of most streams would receive some of them. If the species that can metabolize perchlorates do not form spores, we may be able to give them that capability via genetic engineering by the time terraforming starts.

Currently free oxygen is present in only tiny trace amounts on Mars. Oxygen-fed fires of any kind cannot even burn at partial pressure levels below about 2.25 psi or (71% of Earth partial oxygen pressure). Until oxygen is added, only anaerobic organisms like the botulism bacteria could grow. Also, blue-green algae (or cyanobacteria) would start growing right away in any bodies of water in any that have been cleared of perchlorates as they do not need oxygen to survive. (Most of these organisms would be aquatic.)

Once you can add some significant amounts of oxygen, some kinds of low-oxygen organisms could also grow. The more you add, the more kinds of organisms could grow. Initially, most of these would be aerobic bacteria and single cell eukaryotic algae, as each individual cell is in direct contact with the oxygen dissolved in the water. The minimum level is at least 10% of current oxygen levels or about 0.3 psi in the air. Multicellular animals would probably need at least one-third of current levels or about one psi of oxygen. For small aquatic animals with circulatory systems you might need at least two psi of oxygen, since the animals cells are not in direct contact with the ocean water. For larger animals like fish and amphibians, you would probably need at least 2.5 psi. Animals and fish would not be introduced until there was sufficient plant or algae growth to sustain them.

Most of the early Earth animal models studied for tolerance to low oxygen levels were aquatic, so it is harder to determine the needed levels for land animals. Since many animals live out of water, their lungs would have direct access to the oxygen in the air even though their cells depend on the circulation of blood. All of these obviously have circulatory systems, so any active animals would probably need at least 2.0 psi of oxygen.

Green plants probably require at least 50% of current oxygen levels or about 1.5 psi of oxygen partial pressure compared to today’s sea level oxygen partial pressure of 3.0 psi. Advanced plant roots get less oxygen underground so vascular plants might need 2.0 psi. So the visible greening of Mars would initially be a slow process, speeding up after oxygen levels reached 2.0 psi.

Before anyone should consider starting such a large-scale process, there are a number of provisos and conditions that should be met:

  • The human economy in the Solar System and/or on Mars must be large enough to support the effort.
  • The cost over the construction period would need to be accurately estimated.
  • There must be some level of public or private support.
  • Construction of the equipment can take place over a multi-decade period of time.
  • Replicator systems would greatly reduce cost for the system, so the timing of the start of the operation is important.
  • Sources and levels of heat emission at the fusion and hydrolysis sites need to be identified.
  • Possible damage to equipment from heating could require more active heat control and dispersal systems.
  • The most efficient hydrolysis system adaptable to such a large project should be selected.
  • There should be a priority on reduction in hydrolysis plant mass and wiring mass by good design.
  • Non-heat engine type fusion power conversion methods should be investigated before starting the design process.
  • The North and South Polar layered (water) ice deposits should have a guarantee of protection until sufficient ice coring (and proper storage of the ice cores) is done by climate experts for sufficient past Mars climate information recovery. The minimum or necessary amount of coring will be disputed by climatologists and developers.
  • Once this is done, the polar ice deposits could then be mined directly for the water ice in them if needed. Nitrogen air pressure and oxygen generation come first, large bodies of open water can wait.

The time period for reaching a breathable oxygen/nitrogen atmosphere on Mars can thus be less than it took for some Gothic cathedrals to be built. This major step will eventually allow both microbes and complex, multicellular, eukaryotic life to exist outdoors and in the water on Mars, with no human maintenance needed.

Biological References

Minimum levels of atmospheric oxygen from fossil tree roots imply new plant−oxygen feedback, Fredrik Sønderholm and Christian J. Bjerrum, Geobiology. 2021 May; 19(3): 250–260. Published online 2021 Feb 19. doi: 10.1111/gbi.12435.

Oxygen requirements of the earliest animals, Daniel B. Mills et al., PNAS, February 18, 2014,111 (11) 4168-4172/

Mars Agriculture - Knowledge Gaps for Regolith Preparation Alex Tolley and Doug Loss, Centauri Dreams,11-10-2023.

Note: we are using a new commenting system, which may require you to create a new account.