Quasicrystals have emerged as the third class of solid‐state materials, distinguished from periodic crystals and amorphous solids, which have long‐range order without periodicity exhibiting rotational symmetries that are disallowed for periodic crystals in most cases. To date, more than one hundred stable quasicrystals have been reported, leading to the discovery of many new and exciting phenomena. However, the pace of the discovery of new quasicrystals has lowered in recent years, largely owing to the lack of clear guiding principles for the synthesis of new quasicrystals. Here, it is shown that the discovery of new quasicrystals can be accelerated with a simple machine‐learning workflow. With a list of the chemical compositions of known stable quasicrystals, approximant crystals, and ordinary crystals, a prediction model is trained to solve the three‐class classification task and its predictability compared to the observed phase diagrams of ternary aluminum systems is evaluated. The validation experiments strongly support the superior predictive power of machine learning, with the overall prediction accuracy of the phase prediction task reaching ≈0.728. Furthermore, analyzing the input–output relationships black‐boxed into the model, nontrivial empirical equations interpretable by humans that describe conditions necessary for stable quasicrystal formation are identified.
15 August 2022 in Science and Technology of Advanced Materials: Methods
Materials informatics (MI) research, which is the discovery of new materials through machine learning (ML) using large-scale material data, has attracted considerable attention in recent years. However, in general, the large-scale material data used in MI are biased owing to differences in the targeted material domains. Moreover, most studies on MI have not clearly demonstrated the influence of data bias on ML models. In this study, we clarify the influence of data bias on ML models by combining the concept of the applicability domain and clustering for large-scale experimental property data in the Starrydata2 material database previously developed by our group. The results show that data bias influences the error and reliability of the predictions made by the ML model. The predictions of the ML model within the applicability domain are highly reliable compared to those made outside the domain. This indicates that the material space that can be reliably discovered by the constructed ML model is limited. Nonetheless, we apply the ML model to a large dataset comprising various material classes and find that new materials similar to known materials can be proposed within a limited space. Thus, our findings demonstrate the importance of considering data bias when constructing and evaluating ML models in MI. GRAPHICAL ABSTRACT
Quasicrystals represent a class of ordered materials that have diffraction symmetry forbidden in periodic crystals. Since the first discovery of quasicrystals in 1984, approximately 100 thermodynamically stable quasicrystals have been synthesized. The discovery of new quasicrystals has led to the observation of novel physical phenomena, such as robust quantum criticality, fractal superconductivity, and peculiar long-range magnetic ordering. However, the pace of discovery of new quasicrystals has significantly slowed down, which is attributed to the lack of design principles for exploring new quasicrystals. Here, we demonstrate that machine learning can greatly accelerate the process of material discovery. Our model can predict stable quasicrystalline phases with high accuracy. With this model, we discovered three stable decagonal quasicrystals through an exhaustive screening of more than 1000 ternary aluminum alloy systems.
Quasicrystals are solid-state materials that typically exhibit unique symmetries, such as icosahedral or decagonal diffraction symmetry. They were first discovered in 1984. Over the past four decades of quasicrystal research, around 100 stable quasicrystals have been discovered. In recent years, machine learning has been employed to explore quasicrystals with unique properties inherent to quasiperiodic systems. However, the lack of open data on quasicrystal composition, structure, and physical properties has hindered the widespread use of machine learning in quasicrystal research. This study involves a comprehensive literature review and manual data extraction to develop open datasets consisting of composition, structure types, phase diagrams, and sample fabrication processes for a wide range of stable and metastable quasicrystals and approximant crystals, as well as the temperature-dependent thermal, electrical, and magnetic properties.
23 December 2024 in Science and Technology of Advanced Materials
We introduce our proprietary Materials Informatics (MI) technologies and our chemistry-oriented methodology for exploring new inorganic functional materials. Using machine learning on crystal structure databases, we developed 'Element Reactivity Maps' that displays the presence or the predicted formation probability of compounds for combinations of 80 × 80 × 80 elements. By analysing atomic coordinates with Delaunay tetrahedral decomposition, we established the concept of Delaunay Chemistry. This enabled us to design crystal structures by combining Delaunay tetrahedra of known compounds and to develop the 'Crystal Cluster Simulator' web system. We also developed the Starrydata2 web system to collect large-scale experimental data on material properties from plot images in academic papers. This dataset supported us to select candidate materials for new thermoelectric materials through various data analyses. In large-scale synthesis experiments involving over 7,000 samples, we discovered numerous new phases, including solid solutions of known structures in new combinations of elements. Using sodium metal in synthesis and our proprietary ion diffusion control technologies, we discovered new cage-like compounds by extracting monovalent cations from materials with nano-framework structures, as well as new intercalation compounds. The Element Reactivity Maps were also used to select barrier metals for device electrodes, and an autonomous contact resistance measurement system is under development.
22 April 2025 in Science and Technology of Advanced Materials: Methods