Cracking the Code: How AI Unlocks Secrets of Polycrystalline Materials

Polycrystalline materials are utilized extensively in information equipment, solar cells, and electronic devices; however, these materials can experience a decrease in efficiency due to tiny flaws known as dislocations. To better understand these defects, researchers at Japan’s Nagoya University utilized artificial intelligence. Advanced Materials was the publication that carried out the research.

The production of microscopic crystal flaws due to stress and temperature variations is a big issue with employing polycrystals in industrial applications. A disruption to the normal arrangement of atoms in the lattice can impact electrical conduction and overall performance; these are known as dislocations. It is critical to comprehend the production of these dislocations to lessen the likelihood of failure in devices that employ polycrystalline materials.

Using a novel AI, a group of researchers from Nagoya University, headed by Professor Noritaka Usami and comprising Lecturer Tatsuya Yokoi, Associate Professor Hiroaki Kudo, and others, analyzed picture data of polycrystalline silicon, a material commonly utilized in solar panels. With the use of the AI’s virtual 3D model, the team was able to pinpoint exactly where the material’s performance was being negatively impacted by dislocation clusters. This research may have significant theoretical and practical consequences for our understanding of crystal development and deformation. An important theoretical framework for understanding the behavior of dislocations in materials is the Haasen-Alexander-Sumino (HAS) model. However, according to Usami, they have found dislocations that the Haasen-Alexander-Sumino model failed to capture.

Unexpectedly substantial tensile bond strains along the border of the staircase-like structures that prompted dislocation production were discovered shortly after when the scientists computed the arrangement of the atoms in these structures, which was another surprise.

Electron Microscopy and Theoretical Computations

With the use of the AI’s virtual 3D model, the team was able to pinpoint exactly where the material’s performance was being negatively impacted by dislocation clusters. Researchers employed electron microscopy and theoretical computations to deduce the formation mechanisms of the dislocation clusters after their locations had been identified.
The study, published in Advanced Materials, titled “Multicrystalline Informatics Applied to Multicrystalline Silicon for Unraveling the Microscopic Root Cause of Dislocation Generation,” delves into these questions.

Professor Noritaka Usami oversaw the study of polycrystalline materials, which also involved Lecturer Tatsuya Yokoi, Associate Professor Hiroaki Kudo, and others. They analyzed picture data of polycrystalline silicon, a material commonly used in solar panels, using a novel AI. The team was able to pinpoint the regions where dislocation clusters were impacting the performance of polycrystalline materials thanks to the AI’s creation of a virtual 3D model. Researchers employed electron microscopy and theoretical computations to deduce the formation mechanisms of the dislocation clusters after their locations had been identified.

At the grain boundaries, they discovered staircase-like structures and showed stress distribution in the crystal lattice. During crystal development, these structures seem to induce dislocations. In sum, the study should help in the development of new polycrystalline materials and shed light on the way to creating universal standards for high-performance materials.