電気抵抗ゼロというユニークな性質を持つ超電導材料は，核融合等で期待される未来のエネルギーや作り出したエネルギーを有効に使うための省電力化，強力な磁場を使ったMRIに代表される医療機器のために欠くことの出来ない素材の一つである。近年，機械学習や深層学習によるインフォ

The discovery of new superconductors has a significant impact on the scientific and engineering communities, unraveling interesting physical phenomena and providing unique applications in energy and devices. Superconductors with a high critical temperature are limited to a few families, such as cuprates, iron-based compounds, and hydrides under ultra-high pressure. In traditional studies, the exploration of new superconductors relies on theories, experiments, and simulations. However, recent advances in data science have made machine learning available in a variety of fields, including materials informatics. Utilizing superconductor databases and various regression methods, machine learning has proposed several new superconductors. The chemical descriptors are widely used, and the descriptor of the crystalline structure is being developed for more accurate prediction. In this review, the theoretical and experimental studies for the discovery of new superconductors are explained. The available database and data-driven studies are also shown. Furthermore, after reviewing the recent machine learning studies for the discovery of new superconductors and other materials, future aspects in this field are discussed.

As an emerging style of materials science, we discussed the basic and recent natural language processing technologies, that can be used to collect large experimental dataset for materials informatics. We introduce the classical text-mining for the development of material database, and the approaches to accelerate the automatic data extraction by using recent large language models (LLM). Then we demonstrate how to use the commercial large language models including ChatGPT, by directly asking the LLM how to improve the critical current properties of MgB₂ superconducting wires. By comparing LLM-generated outputs, we analyze word selection and the occurrence of hallucination. Finally, we demonstrate an example to use LLMs effectively, to get inspirations for the development of the best superconducting wire in the history.

The process informatics, which utilizes database constructed from process, computational simulation, measurement and structural analysis data, is a promising method for accelerating materials development. This review describes the currently utilized data science methods and future possibilities in the development of fabrication processes for superconducting materials, such as bulks and thin films, for targeted applications.

The data-driven research has progressed in recent years. Particularly in the field of superconducting materials, there has been active research into the search for new materials, especially by materials informatics (MI). On the other hand, in a superconducting material that has three critical points, the critical current density of the three critical points changes greatly depending on the material manufacturing process. The critical current density properties greatly depend on the material structure, and there are many parameters to obtain the appropriate mictrostructures for various applications. Therefore, these data will be useful to a wide range of users, including materials researchers and even superconducting wire suppliers and a field of application. In this commentary, we provided an overview of data collection, storage, and use, and then presented the progress of some ongoing research and discussed future development possibilities.

Image analysis to identify the phases from microstructural images is an important issue for understanding the mechanism associated with the microstructures of functional polycrystalline materials. In this study, the segmentation ability of the deep learning model and the effect of data augmentation were investigated when applied to ceramic superconducting materials with different compositions than the trained materials.

Collision ESD between charged metallic objects below 1000 V causes electromagnetic interference in electronic equipment and devices. This interference is being observed to intensify at lower charging voltages. The phenomenon was first identified by Masmitsu Honda, but its underlying mechanism remains unclear even today. Originally, the charging voltage and spark length of collision ESD are unknown due to measurement difficulties, making it extremely challenging to establish the relationship between the radiated electric field strength and the spark property in such cases. To clarify the above mentioned electromagnetic phenomenon in collision ESD, in this study, a method of calculating the radiated electric field along with a spark length estimated from the measurement strength of radiated field peak is presented using the spark resistance law developed by Rompe and Weizel. The validity is confirmed by our previous measurement data on radiated electric field due to the collision ESD between charged spherical electrodes with a diameter of 30 mm at charging voltages from 300 to 600 volts, employing an optical field probe with a wideband up to 10 GHz. The estimated spark length is verified by comparing it with the spark lengths based on an empirical Paschen's formula between fixed electrodes and the discharge data from past literature on metal electrodes, revealing the peculiarity effect of the radiated electric field strength caused by “constant breakdown potential gradient” that occurs at charging voltages below 1000 volts.

Frequency and propagation time analyses of the TEV signals were carried out for various distances from patrial discharge source. It was found that the signal immediately after propagation to the TEV sensor contained a high frequency component of 200～300 MHz. Propagation time became slower by increasing distance, and its velocity was close to the velocity of light. Also, this signal intensity reduced with increasing distance.