Researchers combine AI, chemistry to id new materials for energy techs

Researchers at Virginia Tech are bringing artificial intelligence to the search for materials for clean technologies such as fuel cells and carbon capture.

The researchers from Virginia Tech’s College of Engineering have developed what they call a theory-infused neural network (TinNet) approach that integrates deep learning algorithms with well-established theory on catalysts to identify the materials.

Deep learning, a subfield of machine learning, uses algorithms to mimic how the human brain works.

Catalysts are materials that trigger or speed up chemical reactions.

Have you read?
New discovery set to revolutionise ceramic fuel cell technology
Smart metering demands advanced lithium batteries

“About 90% of the products you see today are actually coming from catalysis,” says Hongliang Xin, Virginia Tech associate professor of chemical engineering, who led the research.

“The trick is finding the efficient and robust catalysts for each application, and finding new ones can be difficult. Understanding how catalysts interact with different intermediates and how to control their bond strengths is absolutely the key to designing efficient catalytic processes, and our study provides a tool exactly for that.”

In their report in the journal Nature Communications, the researchers present their work as an alternative to the “appealing” but costly approach of quantum computing, pointing to the orders of magnitude improvements in speed that machine learning offers over traditional computational approaches.

Machine learning algorithms can be helpful because they identify complex patterns in big data sets. But deep learning has limitations, especially when it comes to predicting highly complex chemical interactions – a necessary part of finding materials for a desired function. In these applications, sometimes deep learning fails and it may not be clear why.

The TinNet approach extends its prediction and interpretation capabilities, both of which are crucial in catalyst design.

The researchers intend to make their approach generally accessible to the community for practical use and they anticipate that it will be further developed for renewable energy and decarbonisation technologies.