Advanced Computing in the Age of AI | Thursday, March 28, 2024

Engineers Look to Supercomputing to Find Platinum 2.0 

As a part of President Obama's Material's Genome Initiative, many researchers have set their sights on developing materials the likes of which we've never seen. But engineers at Duke University are using supercomputers to back a materials project that aims to do just the opposite: find a metal that performs like platinum, but at a lower cost.

As a part of President Obama's Material's Genome Initiative, many researchers have set their sights on developing lightweight materials, or unearthing components for the next generation of batteries thanks to the initiative's support of research centers dedicated to accelerating material science and discovery.

In the cases of lightweight materials and batteries, scientists are focused on finding something new - a material with qualities that perform better than what's already being used. But engineers at Duke University are using supercomputers to back a materials project that aims to do just the opposite: find materials that mimic key qualities of those we already use, but come at a much lower cost.

The research, which Duke has highlighted in a new study, comes out of the university's Pratt School of Engineering. And despite its potential for a broad impact, its area of focus is in fact quite small: finding a substitute for platinum.

Despite the focus on a single group of alloys, Duke is doubling down on the metal because of its many applications. Platinum is the go-to material for neutralizing a car's toxic fumes before they escape the exhaust, but it's also essential in producing high octane gasoline, manufacturing plastics and synthetic rubbers, and corralling cancerous tumors.

And although the platinum used for engagement rings doesn't look much like the platinum in a car or factory, one thing they do have in common is a high price.

So if even a single compound in the platinum alloy group is found through this virtual materials research to be less expensive, researchers are saying that numerous industries worldwide could see a monetary benefit in addition to the boon it would provide to the environment.

“We’re looking at the properties of ‘expensium’ and trying to develop ‘cheapium,’” said Stefano Curtarolo, director of Duke’s Center for Materials Genomics. “We’re trying to automate the discovery of new materials and use our system to go further faster.”

To get the job done, Curtarolo and his team are relying on databases and algorithms that have been years in the making. Each works by looking at atomic interactions that build upward to pattern material stability. Already, after almost 40,000 calculations, the team has found 37 new binary alloys in the platinum-group metals, which is already home to osmium, iridium, ruthenium, rhodium, palladium and of course platinum.

For each of these metals, the highlight falls on catalytic properties, or their ability to speed up chemical reactions. But the assets don’t stop there; resistance to corrosion and the ability to get the job done in high-temperature environments make platinum and metals like it ideal for a number of roles that range from electrical components to fuel cells, dentistry and even chemotherapy.

But just because the computer has helped to identify stable platinum-group alloys doesn’t mean that they know how well the materials will manage to fill platinum’s shoes. In fact, studies of Curtarolo’s methodology have found that they provide very little insight into material behavior despite its high accuracy in determining their stability.

“The compounds that we find are almost always possible to create,” said Curtarolo. “However, we don’t always know if they are useful. In other words, there are plenty of needles in the haystack; a few of those needles are gold, but most are worthless iron.”

So now Curtarolo says it’s a matter of turning the materials over to experimentalists to actually produce the alloys and study their physical properties.

“We hope providing a list of targets will help identify new compounds much faster and more cheaply,” said Curtarolo. “Physically going through these potential combinations just to find the targets would take 200 to 300 graduate students five years. As it is, characterizing the targets we identified should keep the experimentalists busy for 20.”

Details of Curtarolo’s research can be found in the December 30 edition of the American Physical Society journal Physics.

EnterpriseAI