Machine Learning method to identify green energy materials

2 mins read

The discovery of proton-conducting oxides, which are crucial in hydrogen fuel cells, is being accelerated by researchers using machine learning.

Machine Learning method to identify green energy materials
Researchers from Kyushu University, in collaboration with Osaka University and the Great Ceramics Centre, are leveraging machine learning to expedite the discovery of proton-conducting oxides vital for hydrogen fuel cells. This innovative approach, detailed in Advanced Energy Materials, promises to revolutionize clean energy production amid escalating climate change concerns. Professor Yoshihiro Yamazaki emphasizes the imperative for efficient hydrogen fuel cells amidst the quest for sustainable energy solutions. Current research underscores the significance of novel electrolyte materials like mica and perovskite oxides, paving the way for enhanced proton conductivity. Conventional detection methods encounter limitations in identifying ideal materials efficiently, prompting the integration of machine learning for data analysis. By forecasting material properties and synthesizing promising candidates, this methodology holds immense potential for advancing oxide fuel cell technologies. The discovery of proton-conducting oxides with unique crystal structures signals a pivotal stride towards realizing environmentally sound societies, setting a precedent for transformative advancements in materials science.

In order to identify and synthesize clean energy materials for use in good oxide fuel cells, researchers at Kyushu University have worked with Osaka University and the Great Ceramics Centre.

These gadgets can produce energy from pure fuels like hydrogen, which doesn’t release carbon dioxide.

Advanced Energy Materials has published a paper titled” Discovery of Unconventional Proton- Conducting Inorganic Solids via Defect-Chemistry-Trained, Interpretable Machine Learning,” which has the ability to hasten the search for materials outside of the energy industry.

Climate change necessitates the production of clean energy.

Scientists have been working on new ways to produce fresh energy to stop the release of more greenhouse gas emissions into the atmosphere. A worthwhile substitute is green hydrogen, a common, light, and very reactive fuel.

But, Professor Yoshihiro Yamazaki of the Department of Materials Science and Technology at Kyushu University stated that in addition to improving how hydrogen is produced, stored, and transported, we even need to increase the efficiency of hydrogen fuel cells in terms of producing electricity.

Study of electrolyte materials

To produce an electrical current, good oxide fuel cells must be able to effectively conduct hydrogen ions through an electrolyte.

The focus of current research into fresh electrolyte materials is on oxides with a mica structure, or crystal arrangement of atoms.

The second proton-conducting oxide was found in a perovskite structure, and scientists are constantly finding high-performing Perovskites.

However, Professor Yamazaki argued that non-perovskite oxides, which also have the ability to conduct protons very efficiently, should be added to the list of good electrolytes.

Difficulties with conventional material detection techniques

There are some drawbacks to using conventional methods to find proton-conducting materials with alternative crystal structures. Smaller amounts of dopant must be added to the base material for an electrolyte to conduct protons.

However, there are a lot of convincing base and doped candidates, each of which has unique atomic and digital characteristics. Thus, it immediately becomes challenging and time-consuming to find the ideal combination that improves proton conductivity.

Data analysis using machine learning

The data was analyzed using machine learning after the researchers calculated the properties of various oxides and dopants. The technology was used to pinpoint the elements that affect a material’s proton conductivity and forecast possible combinations.

The two appealing materials that machine learning had identified were therefore synthesized, and their protons performance was evaluated. Both materials showed proton conductivity in only one experiment.

The materials possible in the future

The first-known proton conductor with a sillenite crystal structure is one of the materials. The conduction paths seen in perovskites are different from those of the different because it has a high-speed proton.

These oxides are now performing poorly as electrolytes. Nonetheless, the team thinks their conductivity can be enhanced with more investigation.

Our framework has the potential to significantly increase the amount of proton-conducting oxides that can be found, which would tremendously hasten the development of strong oxide fuel cells.” According to Professor Yamazaki, it’s a convincing step toward the creation of hygienic societies.

This framework could be applied to other areas of materials science with a few minor changes, probably accelerating the creation of many novel materials.

Verónica Castro-Allende

With a rich background in urban technology and IoT, Veronica brings a unique blend of expertise and innovation to our platform. Her work focuses on integrating advanced engineering solutions into smart city infrastructures, emphasizing sustainable and efficient urban development. Her insightful contributions shed light on the future of urban living, blending technical prowess with practical urban applications

Leave a Reply

Your email address will not be published.

Previous Story

ISCS and PowerStack’s distribution deal

Next Story

EIT InnoEnergy and Demeter announce €500m EBA Strategic Battery Materials Fund

Latest from Technology

Don't Miss