The colonization of the Solar System will not occur without extensive automation using robots and artificial intelligence to handle the construction tasks of permanent bases capable of accommodating numerous Homo sapiens. An illustration of this can be seen in recent work involving the massive production of oxygen on Mars using AI chemistry.
Both Arthur Clarke and Isaac Asimov predicted in the 21st century that the exploration and colonization of the Solar System would only be possible with robots and artificial intelligence. Clarke, in particular, foresaw the World Wide Web as early as the 1960s, providing us with another reason to take his other predictions seriously.
While lunar and space colonies based on the work of Gerard Kitchen O’Neill are undoubtedly dreamt of, Martian colonies seem to capture the imagination the most. However, the journey to Mars and the establishment of a permanent base on the Red Planet are far from simple and will initially be costly until a space industry, similar to that envisioned by O’Neill, is in place with the robotic exploitation of the Moon and asteroids.
Among the challenges for the first Martian colonists was the need for oxygen to breathe. Fortunately, there are regions on Mars with water, and electrolysis could be considered using electric currents generated by solar panels (for reference, the first water electrolysis was carried out in 1800 by two British chemists, William Nicholson and Anthony Carlisle, a few weeks after the invention of the first electric battery by the Italian Alessandro Volta). However, efficient methods are required, not only because the radiation on the surface of Mars is not as intense as that on Earth or the Moon but also due to the necessity for a high yield.
Catalysts Extracted From Martian Ores
Firstly, on our Blue Planet, electrodes with catalysts are employed to reduce the amount of electrical energy required for oxygen production. However, importing catalysts from Earth would be expensive. As a result, a team from the Chinese Academy of Sciences (CAS), under the direction of Professors Luo Yi, Jiang Jun, and Shang Weiwei from the University of Science and Technology of China (USTC), investigated the potential for making effective catalysts from Martian soil.
Indeed, they succeeded in doing so, as demonstrated in an article published in Nature Synthesis. In this publication, they explain their use of artificial intelligence techniques to achieve this outcome.
The researchers initiated the process by obtaining samples from various meteorites of different compositions known to originate from Mars on Earth. These meteorites arrived on Earth after a journey of several million years in space, following ejection from the Red Planet due to a violent asteroid impact on its surface.
Subsequently, the chemical composition of these Martian soil samples was analyzed using the laser-induced breakdown spectroscopy (LIBS) technique.
An AI That Accomplishes 2,000 Years of Work by a Human Chemist in Two Months
Artificial intelligence is fed with the composition of these meteorites and controls various elementary operations in their processing to robotically extract metal hydroxides. These hydroxides are then tested on electrodes.
The same artificial intelligence then conducts simulations of quantum chemistry and molecular dynamics for 30,000 possible combinations of hydroxides. These simulations, along with potential results regarding the efficiency of a catalyst produced from these mixtures, are analyzed by a neural network to predict, even more effectively, the properties of a given mixture of metal hydroxides extracted from lunar soil.
Real tests of these mixtures have been conducted, confirming the identification of viable catalysts and the efficient occurrence of the oxygen-producing electrolysis reaction under average Martian temperatures of approximately -37 °C, at least for one of these catalysts.
The remarkable aspect is that the AI-guided research operations only took two months, whereas a human chemist would have needed to carry out similar experiments for 2,000 years to get the same result.