AI Helps Researchers Discover Catalyst for Green Hydrogen Production

Researchers from the University of Toronto have used AI to generate a “recipe” for an exciting new catalyst needed to produce green hydrogen fuel.

As the effects of climate change begin to become more apparent in our everyday lives, research like this could open the door to green hydrogen fuel that could be used for everything from transportation to residential and commercial heating.

While the entire process of creating hydrogen fuel is quite complicated, it can be summarized more simply.

Scientists take water and pass electricity from renewable sources between two pieces of metal called electrodes that are submerged in the water. These electrodes are coated in a catalyst that speed up the process of splitting the water into its two parts – hydrogen gas and oxygen gas. This hydrogen gas is then taken and can be used for fuel.

Until now, iridium oxide was the most widely used catalyst that could withstand the harsh acidic conditions in this reaction. Sadly, iridium is extremely scarce and expensive. This makes it an unsustainable source for large-scale hydrogen production. Ruthenium-based catalysts are more abundant and less expensive than iridium, but they suffer from instability due to the overoxidation of ruthenium atoms during the reaction.

Therefore, the scientists at the University of Toronto endeavored to use AI resources to solve this problem. The team created an AI program to speed up the search for an optimal alloy combination that would act as a catalyst in the water-splitting reaction. This program analyzed over 36,000 different metal oxide combinations through virtual simulations. Traditionally, such a search would require trial and error in the lab.

The Canadian Light Source at the University of Saskatchewan allows scientists to look at things that we can’t see with our eyes or regular microscopes.

“We’re talking about hundreds of millions or billions of alloy candidates, and one of them could be the right answer,” Jehad Abed, who was part of the team that developed the AI program, said in a blog post about the research.

After letting the AI program chew through the 36,000 different metal oxide combinations, the program recommended a combination of ruthenium, chromium, and titanium. Aben then tested the program’s top candidate in the lab to see if the program was accurate. Then, the team used the Canadian Light Source (CLS) at the University of Saskatchewan to help researchers understand how the new catalyst works.

Think of the CLS as a super-powerful microscope that uses bright X-rays to look at things we can’t see with our eyes or regular microscopes. The scientists shined these bright X-ray lights on their new catalyst while the catalyst was working to split the water. This allowed them to see how the atoms in the catalyst were arranged and how they moved during the reaction. Specifically, they wanted to ensure the ruthenium didn’t dissolve, which can often happen due to overoxidation.

After this process, the scientists found that their new catalyst was better at keeping the ruthenium from dissolving than other candidates. Additionally, they saw that this catalyst was less likely to break apart oxygen from its own structure, thereby helping it last longer. What’s more, they witnessed some pretty incredible results.

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

This research is considered a major success, but scientists still have a long way to go before the ruthenium, chromium, and titanium alloy can be used in large-scale hydrogen production. The scientists stated that they would need a lot of testing to ensure it would last under “real world” conditions.

The work performed here is an example of how AI can offer faster routes to finding answers to the complex questions surrounding the climate crisis.