#Japan #Tokyo #NIMS #Element Map #New Materials

Discovering New Materials through Machine Learning

In a groundbreaking achievement, the National Institute for Materials Science (NIMS) has collaborated with top institutions such as the University of Tokyo, the National Institute of Advanced Industrial Science and Technology, Tohoku University, and Kyoto Institute of Technology to develop and publicly release an innovative tool, the "Elemental Reactivity Map." This advanced resource aims to revolutionize the way researchers discover new materials by harnessing the power of machine learning.

The Problem at Hand

Traditionally, synthesizing inorganic materials involves combining various elements. However, the challenge arises when trying to create new compounds that have never been synthesized before. If a new material possesses unique properties or useful functions, it could lead to exciting advancements in practical applications. Unfortunately, many potential combinations are not found in existing crystal structure databases, as they have simply never been tested due to previous unsuccessful attempts. Therefore, predicting which combinations might yield new materials is essential for efficient material discovery.

Key Innovations

The team at NIMS developed a comprehensive set of 80 grids known as the "Elemental Reactivity Map," which visually represents the potential for creating substances using combinations of up to three different elements. By employing machine learning techniques based on a dataset of over 30,000 inorganic compounds and their crystal structures, they have identified more than 3,000 promising new elemental combinations.

These maps are now publicly available as an interactive web system, allowing researchers and interested individuals alike to access the wealth of information freely. This accessibility opens up the potential for widespread collaboration and exploration in material science.

To validate the predictions made by the maps, the team utilized a database of complex crystal and solid solution structures. They found that combinations with high reactivity scores (above 0.95) were 17 times more likely to correspond to known compounds compared to those with low reactivity scores (below 0.05). This significant finding supports the credibility of the reactivity score system.

Furthermore, the research uncovered over 3,000 combinations with high reactivity scores that have not yet been documented in experimental databases, hinting at the existence of new materials waiting to be discovered. These combinations may well represent hidden gems in the search for innovative materials.

For practical applications, the Elemental Reactivity Map has already facilitated the identification of numerous new materials. Notably, it has led to discoveries in B20 structured alloys like Co(Al,Ge), known for their magnetic skyrmion and thermoelectric properties.

By merging cutting-edge technology with scientific inquiry, NIMS and its collaborators are paving the way for revolutionary advances in material science. The implications of this research are vast, potentially influencing various industries and research fields as they seek to innovate and implement new materials.

For more detailed information about this initiative, please refer to the official press release here.