AI chemist finds molecule to make oxygen on Mars after sifting through millions

A spectacular example of gullies on Mars, at roughly 71 degrees latitude in the southern hemisphere.
Gullies on Mars, at roughly 71 degrees latitude in the southern hemisphere. (Image credit: NASA/JPL-Caltech/University of Arizona)

Using meteorites from Mars, an AI-powered robot chemist synthesized compounds that could be used to generate oxygen from water, scientists announced on Monday (Nov. 13).

Potential future crewed missions to Mars will need oxygen — not just for astronauts to breathe, but also for use as rocket propellant. One key way to make such missions cost-effective in the long run is to use resources that already exist on the Red Planet to create the oxygen. That'd be much easier than lugging a bunch of oxygen, and oxygen-producing materials, all the way from Earth.

The idea is promising because Mars does possess significant reserves of frozen water ice — because water is made of hydrogen and oxygen, scientists have been looking for ways to harvest the latter element from those Martian reserves. In particular, compounds known as catalysts are capable of spurring chemical reactions that "split" water molecules to generate oxygen and hydrogen gas.

Related: Mars ice deposits could pave the way for human exploration

In a new study, researchers experimented with an AI chemist to produce some of those water-splitting catalysts — most importantly, these tests were conducted with materials found on Mars. The team focused on five different categories of Martian meteorites, which are rocks that crashed down on Earth after cosmic impacts blasted them off the Red Planet.

The AI chemist used a robot arm to collect samples from the Martian meteorites, then it employed a laser to scan the ore. From there, it calculated more than 3.7 million molecules it could make from six different metallic elements in the rocks — iron, nickel, manganese, magnesium, aluminum and calcium. 

Within six weeks, without any human intervention, the AI chemist selected, synthesized and tested 243 of those different molecules. The best catalyst the robot found could split water at minus 34.6 degrees F (minus 37 degrees C), the kind of cold temperature found on none other than Mars. 

"When I was a boy, I dreamed of interstellar exploration," Jun Jiang, co-senior author of the study and a scientist at the University of Science and Technology of China in Hefei. told "So when we finally saw that the catalysts made by the robot could actually produce oxygen by splitting water molecules, I felt like my dream was coming true. I even started to imagine that I, myself, will live on Mars in the future."

The researchers estimate it would have taken a human scientist something like 2,000 years to find that "best" catalyst using conventional trial-and-error techniques. Still, Jiang noted that, although these findings suggest AI can be very helpful in science, it "at the same time needs the guidance of human scientists. The robot AI chemist is smart only if we taught it to do something."

The scientists now aim to see if their AI chemist can operate under Martian conditions other than temperature, "in which the atmospheric composition, air density, humidity, gravity and so on are so different than those on Earth," Jiang said.

The researchers detailed their findings online on Monday (Nov. 13) in the journal Nature Synthesis. 

Join our Space Forums to keep talking space on the latest missions, night sky and more! And if you have a news tip, correction or comment, let us know at:

Charles Q. Choi
Contributing Writer

Charles Q. Choi is a contributing writer for and Live Science. He covers all things human origins and astronomy as well as physics, animals and general science topics. Charles has a Master of Arts degree from the University of Missouri-Columbia, School of Journalism and a Bachelor of Arts degree from the University of South Florida. Charles has visited every continent on Earth, drinking rancid yak butter tea in Lhasa, snorkeling with sea lions in the Galapagos and even climbing an iceberg in Antarctica. Visit him at

  • JoeReal
    We really need to extract oxygen from oxides in the rocks and soils of Mars instead of split it from the scarce water. What they found out is a catalyst found in the Mars meteorite that could split water. We already have efficient water splitting catalysts here on earth that can be easily transported to Mars as it doesn't require a lot of material, and we don't need to send a catalyst manufacturing factory to Mars, just send our best catalyst if the only intention is to split the very limited amount of water on Mars! Water on Mars is a finite very severely limited resource and must be conserved for other purposes when it is inhabited by humans, such as for farming or food production. The best solution is to develop a method to extract the oxygen from rocks and soils as they contain a lot more oxygen. So this finding isn't really helpful but the newly found catalyst might be better than the ones we have here on earth for splitting water but we'll have to get raw materials from Mars if we can't synthesize it here, and if we use it on Mars, we may need to bring a factory to process it there, which will be more expensive than using the best catalysts we already have.
  • JoeReal
    let's hope the AI Chemist can develop an efficient process to extract oxygen from oxide minerals and soils of Mars.
  • michael1901n
    I'm actually excited for Mars now and how we can colonize it in the future! Although i am a bit worried about how if people are going to colonize Mars, it will become a bit more crowded. These discoveries get us closer and closer to the future, to Mars we go! Oh and the metals and how we can extract air from it is really going a mile away! I am excited to go to Mars and explore the wonders that is space...
  • newtons_laws
    Although the article makes no mention of it, in order to split water into oxygen and hydrogen you of course need an input of energy, since when hydrogen and oxygen combine to form water the reaction is exothermic (i.e energy is released), so by conservation of energy you have to put energy back in to separate them. An uninformed person reading the article might assume that the catalyst splits water into hydrogen and oxygen all on its own! However if you go to the journal Nature Synthesis that is linked in the article it is clear that the researchers are proposing that the catalyst is used in an electrochemical workstation that uses electrolysis powered by solar panels to split the water, the catalyst just makes the electrolysis process more efficient than with no catalyst.
  • Classical Motion
    Electrolysis is a current flux, a heat flu, to break water. Only certain rates of that flux is doing the work.

    But if we knew the right rate........and could mimic it with an EM pulse, we might separate it with little effort.

    I believe that in the future, chemistry will be electronically controlled........not with heat flux and pressure.

    I think in the future, atoms and molecules be be looked at in a different manner. For they are nature's electronic circuits.
  • billslugg
    There is no free lunch, catalysts just speed up the waiter.
    Put a conventional electrode into water, apply a voltage, a very small amount of current will flow and a small amount of water is electrolyzed. Each molecule of oxygen takes a certain amount of energy to dislodge.
    Make the electrodes out of a catalyst and the reaction speeds up dramatically. The current will also rise. The amount of energy per oxygen molecule remains the same.
  • Stahrrman
    Wow! We're not even there yet and we're already devising ways to deplete the resources. Now that's progress! ;)
  • Classical Motion
    We(human beings) come with nothing. We must hunt and consume resources and materials and manufacture tools and structures to exist. And to make life even partly comfortable we must convert materials into needed products at high in-efficiencies.....usually with heat. This has given us a modern life.....via electricity.

    Lucky for us, this planet can take our waste and recycle it into the resources we need. And we are not alone in this process, all life forms have and need this dynamic.

    All lifeforms consume and deplete the environment around them. They also excrete waste. Life also consumes energy, converts it and excretes it. And comparing other energy densities, life has energy density much higher than dead matter......thru-out the universe.

    What's the chances of you finding such an environment? Another world like that?

    What's the chances of A.I. Terra-forming such an environment with stability or balance?

    I believe from what I have read that we have plenty of resources for hundreds of years. And for this climate worry, we will know within ten years if these CO2 theories and predictions are true.

    We should realize by now that yearly increases in CO2 is going to be a reality. Only great war or great disease/famine can prevent it.

    And with our current climate predictions, we haven't the needed time or resources to colonize anything.

    It would be easier to colonize Greenland or Antarctica than Mars. Both have valuable resources. It would be easier to farm under a mile of ice, than farming on Mars. We could do that.
  • Classical Motion
    What if it turns out that water is the result of a bio-process and chemically induced water(molecule, not condensed) is a rarity? Maybe water did not might have been made.

    In the last 20 years we have found new life forms that over double the present known life forms in bio mass of this planet. Many suspect much more to be found. This was done with environmental DNA and most have not been found naturally and studied, grown and analyzed. We know nothing of what they do OR where they fit into the environment. Or how they fit in the recycle scheme.

    We know so little. We have only a glint.