Computer simulations quickly discover an alloy of ruthenium, chromium, and titanium as a more durable, efficient, and greener catalyst.

Researchers at the University of Toronto are using artificial intelligence to accelerate scientific breakthroughs in the search for sustainable energy – by using the Canadian Light Source (CLS) at the University of Saskatchewan to confirm an AI-generated formula for a new catalyst could potentially produce hydrogen fuel more efficiently.

Green hydrogen is produced by energizing two metal electrodes submerged in water with renewable electricity. The submerged catalysts split the water molecules, releasing oxygen and hydrogen gases. However, this is currently an energy-intensive process and often utilizes catalysts comprising metals that are rare and costly.

A research team at the University of Saskatchewan recently confirmed that an AI-generated catalyst recipe offers a more efficient method for producing hydrogen fuel.

Researchers have been searching for the optimal combination of materials to act as a more affordable catalyst to create the reaction, a task that can take years of trial-and-error and testing of millions of metal combinations in a lab – with AI simulations, identifying the top hydrogen can now be done in days.

"We're talking about hundreds of millions or billions of alloy candidates, and one of them could be the right answer," said Jehad Abed, a member of the team that developed the AI to significantly speed up this search.

The AI program analyzed over 36,000 metal oxide combinations through virtual simulations to determine the most promising options; Abed then tested the top candidate in the lab.

"The computer's recommended alloy performed 20 times better than our benchmark metal in terms of stability and durability," he said. "It lasted a long time and worked efficiently."

An alloy made from ruthenium, one of the six platinum group metals known for their catalytic properties, mixed chromium and titanium, was a clear winner.

"The computer was right about this alloy being more effective and stable. That was a breakthrough because it shows that this method for finding better catalysts is working," said Abed. "What would take a person years to test, the computer can simulate in a matter of days."

The team then utilized the CLS's ultra-bright X-rays to assess the catalyst's performance. The CLS is a synchrotron, which produces different types of light to study the structural and chemical properties of materials by recording the ways light interacts with the material's individual molecules.

"What we needed to do is use that very bright light at 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 that we put in," Abed explained.

The researchers also utilized the Advanced Photon Source at Argonne National Laboratory in Chicago.

"After identifying the candidate, we can now go to the lab, make the candidate material and then test it in a real device," he said, with stations at the CLS able to shine a very bright light into the catalyst material during the reaction itself. "This allowed us to not only create the most efficient catalyst, but also understand important things about the underlying mechanism at play. Understanding this was very important for our study."

While the AI program marks a breakthrough in the method for finding better catalysts, any material discovered will still require extensive testing to ensure it performs well under real-world conditions.