AI-powered team finds cheaper way to produce green hydrogen

Researchers at the University of Toronto are using artificial intelligence to accelerate scientific breakthroughs in the search for sustainable energy. They used the Canadian Light Source (CLS) at the University of Saskatchewan (USask) to confirm that an AI-generated “recipe” for a new catalyst offered a more efficient way to produce hydrogen as a fuel.

To create green hydrogen, electricity generated from renewable resources must be passed between two pieces of metal in water. This causes oxygen and hydrogen to be released. The problem with this process is that it currently requires a lot of electricity and the metals used are rare and expensive.

Researchers are looking for the alloy or combination of metals that would serve as a catalyst to make this reaction more efficient and affordable. Traditionally, this search would involve trial and error in the lab, but when it comes to finding the needle in a haystack, this approach takes too long.

“We’re talking about hundreds of millions or billions of candidate alloys, and one of them could be the right answer,” said Jehad Abed. He was part of a team that developed a computer program to dramatically speed up the search.

The results are published in the Journal of the American Chemical SocietyAt the time of this project, Abed was a doctoral student under Edward Sargent at the University of Toronto and working alongside scientists at Carnegie Mellon University.

The team’s AI program considered more than 36,000 different combinations of metal oxides and ran virtual simulations to assess which combination of ingredients might work best. Abed then tested the program’s top candidate in the lab to see if its predictions were accurate.

The team used the CLS’s ultra-bright X-rays to analyze how the catalyst behaves during a reaction. “We needed to use this very bright light from the Canadian Light Source to shine it on our material and see how the atomic arrangements would change and respond to the amount of electricity we put into it,” Abed said. The researchers also used the Advanced Photon Source at Argonne National Laboratory in Chicago.

The alloy, a combination of ruthenium, chromium and titanium in specific proportions, was a clear winner, Abed said.

“The alloy recommended by the computer performed 20 times better than our reference metal in terms of stability and durability,” he said. “It lasted a long time and worked efficiently.”

While the AI ​​program developed by Jehad and his colleagues shows great promise, the material itself still needs to undergo extensive testing to ensure it will last under “real-world” conditions.

“The computer was right that this alloy was more efficient and more stable. This was a major breakthrough because it proves that this method of finding better catalysts works,” Abed said. “What would take a person years to test, the computer can simulate in a matter of days.”

Researchers hope AI will offer a faster way to find the answers we need to make green energy practical for widespread use.