Crystalline materials play a crucial role in various industries, including semiconductors, pharmaceuticals, photovoltaics, and catalysts. As scientists continue to design new materials to address emerging challenges, the need for precise identification methods becomes increasingly important. Currently, powder X-ray diffraction is widely used for this purpose, but it becomes complex when dealing with multiphase samples containing different types of crystals. To expedite the identification process, researchers have turned to innovative data-driven methods, such as machine learning. Now, a new study led by researchers from Tokyo University of Science (TUS) proposes a machine learning “binary classifier” model that can identify the presence of icosahedral quasicrystal (i-QC) phases from multiphase powder X-ray diffraction patterns.
Identifying the various phases present in multiphase samples has traditionally relied on the expertise of scientists, making the process time-consuming. However, the development of a machine learning model specifically designed for identifying i-QC phases brings about a promising solution. The researchers developed a “binary classifier” model using 80 types of convolutional neural networks and trained it using synthetic multiphase X-ray diffraction patterns representing i-QC phases. The model’s performance was assessed using both synthetic patterns and a database of actual patterns, and it achieved an impressive prediction accuracy of over 92%.
Accurate Identification of Unknown Phases
The proposed machine learning model not only demonstrated its ability to accurately identify known i-QC phases but also successfully identified an unknown i-QC phase within multiphase Al-Si-Ru alloys. This finding was confirmed through the analysis of the material’s microstructure and composition using transmission electron microscopy. Importantly, the model was able to identify the i-QC phase even when it was not the most prominent component in the mixture. The researchers believe that this approach can also be extended to the identification of new decagonal and dodecagonal quasicrystals and other types of crystalline materials.
The success of this deep learning model in detecting unknown quasicrystalline phases in multiphase samples has significant implications. By accelerating the process of phase identification, this model opens up new possibilities in the field of materials science. The accurate and efficient identification of new phases can lead to advancements in materials design and development. Furthermore, the model’s applicability to various types of crystalline materials broadens its potential impact on diverse industries. This innovative approach to phase identification lays a foundation for further advancements in the study of crystalline materials found in mesoporous silica, minerals, alloys, and liquid crystals.
The development of a machine learning “binary classifier” model capable of identifying icosahedral quasicrystal phases in multiphase samples represents a significant breakthrough in the field of materials science. By utilizing synthetic X-ray diffraction patterns and convolutional neural networks, the model achieved a high prediction accuracy, even for unknown phases. This innovative approach has the potential to greatly expedite the identification of new phases and enhance materials design and development processes. The successful identification of unknown i-QC phases paves the way for future applications in various types of crystalline materials. As researchers continue to push the boundaries of machine learning in materials science, the possibilities for advancements and discoveries in this field are vast.