Machine learning accelerates discovery of high-performance metal oxide catalysts

Machine Learning


This article has been reviewed in accordance with Science X's editorial processes and policies. The editors have highlighted the following attributes while ensuring the authenticity of the content:

Fact-checked

Peer-reviewed publications

A trusted source

Proofread


Workflow of the ML-based analytical process adopted to explore multicomponent ORR catalysts under alkaline conditions. credit: Journal of Materials Chemistry A (2024). DOI: 10.1039/D4TA01884B

× close


Workflow of an ML-based analytical process used to explore multicomponent ORR catalysts under alkaline conditions. credit: Materials Chemistry Journal A (2024). Posting date: 10.1039/D4TA01884B

Researchers have harnessed the power of artificial intelligence to make significant advances in the discovery and optimization of multicomponent metal oxide electrocatalysts for oxygen reduction reactions (ORR).

This breakthrough has the potential to revolutionize the efficiency and affordability of renewable energy technologies such as hydrogen fuel cells and batteries, paving the way to a sustainable energy future.

For details of the survey results, Journal of Materials Chemistry A April 23, 2024.

In this study, the researchers analyzed 7,798 different metal oxide ORR catalysts from a high-throughput experiment. These catalysts, containing elements such as nickel, iron, manganese, magnesium, calcium, lanthanum, yttrium, and indium, were tested at a range of potentials and their performance was evaluated.

The researchers used the XGBoost machine learning technique to build a predictive model to identify potential new compositions without the need for exhaustive experimental testing.

The study found that a large number of itinerant electrons and high configurational entropy are important features to achieve high current densities in ORR. At a current density of 0.8 VRHE, the Mn-Ca-La, Mn-Ca-Y, and Mn-Mg-Ca ternary systems showed great potential for hydrogen fuel cell applications. 0.63 VRHE identified Mn-Fe-X (X = Ni, La, Ca, Y) and Mn-Ni-X (X = Ca, Mg, La, Y) systems as promising candidates for hydrogen peroxide production. I did.


(ab) (a) Comparison of R2 (b) RMSE between models built by ANN, XGBoost, and LightGBM on the training and test sets. (c–d) Comparison of experimental and predicted values ​​by XGBoost on (c) the training set and (d) the test set. RMSE is in lg(μA cm-2). credit: Journal of Materials Chemistry A (2024). Posting date: 10.1039/D4TA01884B

× close


(a–b) (a) Comparison of R2 (b) RMSE between models built on training and test sets by ANN, XGBoost, and LightGBM. (c–d) Comparison of experimental and predicted values ​​by his XGBoost on (c) training set and (d) test set. The unit of RMSE is lg(µA cm-2). credit: Materials Chemistry Journal A (2024). DOI: 10.1039/D4TA01884B

“Our innovative approach using machine learning accelerates the design and optimization of multicomponent catalysts, saving considerable time and resources,” said Xue Jia, an assistant professor at the Institute for Advanced Study in Materials Science and one of the study's co-authors.

“The efficient identification of high-performance catalyst compositions provides a proven, transformative approach that could lead to major advances in sustainable energy technologies.

Enhanced catalysts can improve the efficiency and reduce costs of renewable energy technologies, fostering their widespread adoption and reducing dependence on fossil fuels. More efficient energy storage systems reduce overall costs, make renewable energy more accessible, and help protect the environment.

The successful application of machine learning in this study sets a precedent for future research and could lead to breakthroughs in various scientific fields. The improved ORR catalyst could also enhance the production of hydrogen peroxide, which is widely used in disinfection and industrial processes, benefiting public health and safety.

“This work highlights the incredible potential of artificial intelligence to accelerate catalyst design and materials discovery,” Jia added. “We hope that our findings will enable future breakthroughs in sustainable energy technologies, which are essential to addressing the world's energy challenges.”

For more information:
Xue Jia et al, Machine learning enables exploration of multicomponent metal oxides as oxygen reduction catalysts in alkaline media, Materials Chemistry Journal A (2024). DOI: 10.1039/D4TA01884B

Magazine information:
Journal of Materials Chemistry A



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *