Robotic AI chemist could make oxygen on Mars

To develop a catalyst for those reactions, Weiwei Shang, Jun Jiang, Yi Luo and their colleagues at the University of Science and Technology of China made a machine with a robotic arm that used laser spectroscopy to analyze the composition of five meteorite samples. Its AI “brain” used theoretical simulations to suggest over 3.7 million potential catalysts based on six metals in the ores, then used machine learning to predict the 243 that would best split water. The robot synthesized and tested those candidates to come up with the best catalyst. All of this took 6 weeks; a human chemist would have taken 2,000 years.

The acids and alkali components needed to treat the meteorites could in principle be made from Martian brine, Jiang says. “For every square meter of Martian material, our robot could make nearly 60 grams of oxygen per hour.” Humans need roughly 850 g of oxygen per day on average, so the AI chemist would need only 15 h of solar power to produce enough oxygen for human survival.