Thermoelectric materials are envisaged to be used for power generation applications, ranging from power sources of myriad IoT sensors to energy-saving waste heat power generation, and also solid-state cooling. In this review I present an overview of the development strategies of the state-of-the-art Mg-Sb based materials which can exhibit superior thermoelectric properties to the long-time champion rare Bi₂Te₃-type materials. Equally importantly is also covering the recent developments regarding application-related technology of these new materials which is critical for realising real-world applications. Namely, development of novel principle active electrodes and diffusion barrier layer, and improved design of thermoelectric devices, culminating in the recent AI neural network TEGNet which enables 10,000 times faster optimization compared to conventional methods.

For the studies of thermoelectricity and thermal transport such as Seebeck coefficient and thermal conductivity, Boltzmann transport theory is quite often used which assumes the semi-classical equation of motion of electrons. However, it is desired that we go beyond the Boltzmann equation to evaluate physical quantities, and develop new possibilities including new thermoelectric materials by understanding the thermoelectric phenomena in the framework of microscopic theory such as linear-response theory. In this article, we explain the recent developments including our theoretical studies.

Materials that exhibit negative thermal expansion (NTE)—that is, they contract when heated—have attracted increasing attention as thermal‑expansion compensators that can improve dimensional precision of components and mitigate failures caused by thermal stress and thermal strain in systems and devices, in line with advances in industrial technology. This article summarizes representative NTE materials and the mechanisms underlying their behavior. It also introduces recent research trends, focusing on the authors’ own studies, including material development and applications to thermal‑expansion control.

As an innovative alternative to the gas stripper, which has been a major obstacle to downsizing conventional accelerator mass spectrometry (AMS) systems, we aimed to apply a "crystal surface stripper method" to a downsized AMS system. Using a prototype system with a maximum acceleration voltage of 45 kV, we experimentally evaluated the charge exchange and molecular dissociation characteristics of carbon negative ions on potassium chloride, tin telluride, and gold surfaces. The results showed that the observed charge state distributions were in good agreement with a resonant charge transfer model. On the other hand, the dissociation efficiency of interfering molecules did not reach the required suppression level (10⁻¹²) even at 45 keV, revealing the need for further improvement. This study highlights the remaining challenges for realizing next-generation downsized AMS systems and establishes an important foundation for future development.

We have developed a non-destructive measurement technology for concrete structures that can be used outdoors to address salt damage, one of the three major deterioration factors for many infra-concrete structures such as bridges. This technology involves irradiating the target object with neutrons and measuring the prompt gamma rays produced by the reaction with the elements present. Using neutron-induced prompt gamma ray analysis, a quantitative method for elements, it measures the chloride ion concentration distribution in the depth direction necessary for inspecting salt damage in concrete bridges. It detects the chloride ion concentration of 1.2 kg/m³, which steel begins to corrode, and is a new non-destructive measurement technology that enables complete non-destructive measurement without damaging the structure.

In recent years, magnetic skyrmions, vortex-like structures of electron spins, have been extensively studied as candidates for information carriers. In particular, recent research has identified novel materials which host nanometric skyrmion with a few nanometers in diameter, driven by a novel mechanism based on itinerant-electron-mediated interactions. In this article, we introduce the latest achievements in the exploration of skyrmion-hosting materials and discuss the prospects for next-generation magnetic memory devices utilizing skyrmions.

Thin-film growth optimization is costly and prone to experimental failures, and has traditionally relied on expert intuition. This tutorial introduces Bayesian optimization (BO) and physics-informed Bayesian optimization (PIBO) as key techniques for closed-loop autonomous synthesis. We summarize practical principles and recent advances for stably and efficiently iterating the loop of condition proposal, growth, characterization, and machine learning model update, illustrated with the authors’ case studies. We also discuss how data accumulated through autonomous thin-film growth can be leveraged to extract growth rules and actionable process knowledge.