狐狸视频

U of T Engineering researchers use machine learning to design smarter industrial catalysts

""
A computer algorithm developed at U of T analyzes thousands of possible geometric configurations of five different elements in order to predict which ones will have the best performance as industrial catalysts (Image courtesy of Zhuole Lu)

A team of researchers from the 狐狸视频鈥檚 Faculty of Applied Science & Engineering is leveraging machine learning to enhance the manufacture of everyday items.

The researchers created a machine learning algorithm that efficiently sifts through thousands of possible metal element configurations in order to design better industrial catalysts, which, in turn, improve chemical reactions.

The results of the research were recently .

鈥淐atalysts speed up chemical reactions, making manufacturing more efficient and less costly,鈥 says Chandra Veer Singh, an associate professor in the department of materials science and engineering. 鈥淭hey are used to produce everything from plastics to industrial fertilizers and researchers around the world are constantly trying to improve their performance.

鈥淎lloying is a key strategy in this field.鈥

An alloy is a combination of two or more metal elements such as platinum or iridium. Many catalysts are binary alloys that contain two metals, or tertiary alloys that contain three. But in 2018, a team of American researchers developed a new synthesis technique that allows an alloy of five different metals to be used as a catalyst.

鈥淭hese alloys are referred to as high-entropy alloys (HEA) because the large number of elements creates a high level of entropy, which is a measure of amount of disorder in the material,鈥 says Zhuole Lu, a graduate student who is pursuing a master鈥檚 degree in materials engineering. 鈥淪ome HEAs may turn out to be excellent catalysts, but the challenge is that there are just too many possible combinations.鈥

For example, if a researcher has 30 metals to choose from, there are about 400 ways to create a binary alloy and about 4,000 ways to create a tertiary alloy. But there are more than 140,000 ways to create a five-element HEA.

Those numbers become even larger when geometric effects are taken into account. Even for a single five-element combination, there are millions of different ways those same elements could be arranged on an alloy surface.

Creating and testing all the potential combinations and arrangements would be a costly and time-consuming way to find new catalysts. So, Singh, Lu and post-doctoral researcher Zhi Wen Chen came up with a new approach.

The team鈥檚 machine learning algorithm 鈥 known as a neural network 鈥 can realistically simulate the properties of a given HEA without having to actually produce it in a lab.

鈥淲hat makes this work special is its inclusion of structural effects,鈥 says Chen. 鈥淵ou can imagine the metal atoms in the HEA as plastic balls in a children鈥檚 ball pit. Current machine learning models treat the alloy as though it had a flat surface. But we know that the top of the ball pit is bumpy, and that bumpiness can have a big impact on the catalyst鈥檚 properties. Our model takes that into account.鈥

The team validated the model by predicting the catalytic performance of a five-element HEA that contained the metals iridium, palladium, platinum, rhodium and ruthenium. They then compared the predictions of the machine learning model with experimental measurements from five different research groups. The results matched surprisingly well.

In the future, the team hopes to expand the approach, creating a model that can simulate different combinations of elements as well as different geometries. Such an algorithm could quickly point researchers toward the most promising catalyst candidates.

鈥淏y simultaneously accounting for different metals and different structures, we believe our tool will help materials scientists rapidly screen for the best new catalysts, while also enabling a deeper understanding of HEAs and alloy catalysts in general,鈥 says Singh.

The Bulletin Brief logo

Subscribe to The Bulletin Brief

Engineering