A team of researchers from Hefei, China, has made a significant breakthrough in developing a catalyst using materials found on Mars. The main goal is to produce oxygen on Mars, which requires a catalyst. Transporting heavy equipment to Mars is expensive and time-consuming, making it crucial to manufacture necessary devices directly on the planet using available resources.
The research team employed artificial intelligence (AI) to create a catalyst under laboratory conditions. The AI was trained specifically for this purpose and executed the entire process with minimal human intervention. This included developing a catalyst formula, pre-treating the ore, synthesizing the catalyst, and testing it post-production.
Researchers estimate that the automated process saved about 2,000 years of human labor. This figure comes from the study detailing their work. Initially, the team developed an AI chemist robot capable of performing these tasks. The robot was based on previous AI chemists with limited abilities to read synthetic chemistry literature and assess the effectiveness of various chemical compounds for different tasks. Once the model was created, it was enhanced with additional data.
The researchers selected five different rocks from the Martian surface, predicting around 3.8 million possible combinations of elements in these rocks. They let the AI choose the required combination, supporting it with 30,000 other theoretical datasets and results from 243 experiments. The AI identified a “polymetallic” material composed of manganese, iron, nickel, magnesium, aluminum, and calcium.
Following this, the AI built a prototype of the catalyst. It used a robotic arm to take physical samples from meteorites dissolved in hydrochloric acid. The synthesis process involved extreme methods, such as centrifuging samples at 7,500 g for five minutes to separate necessary materials. The resulting material was then dried.
Once a portion of the material was synthesized, the research team conducted a reduction process under Mars-like conditions. The material demonstrated excellent performance, comparable to existing catalysts.
The most remarkable aspect of this process is that it operated with almost no human intervention, showcasing the current capabilities of AI. Whether the developed catalyst will be used in the future remains to be seen, but the fact that AI can solve complex technical problems and exhibit creativity is impressive.