The growth of the information society has made the increasing energy consumption associated with computation and data storage a major societal challenge. To improve computational performance while reducing power consumption, hardware adapted to different computing principles and operating environments is being developed, including analog in-memory computing for AI and cryogenic control circuits for quantum computers. As these hardware technologies advance, it is becoming increasingly important to explore materials and physical properties suited to the functions and operating environments of components ranging from computing elements to peripheral circuits.
Our research focuses on phase transitions characterized by large changes in electrical conductivity, such as metal–insulator transitions and superconductivity, from the perspective of developing material functionalities. In particular, we use rapid heating and quenching to bypass equilibration and generate metastable phases, thereby exploring methods for nonvolatile control of physical properties. Thermal quenching has conventionally been used in alloys, glasses, and related materials to suppress structural transformations that require long-range atomic diffusion. We are extending this concept to electronic phase transitions that occur without such long-range atomic diffusion. As one example, we discovered a metastable superconducting phase by using thermal quenching to avoid a transition into a competing electronic order. We further demonstrated nonvolatile and reversible switching of the superconducting state through its creation and erasure using current pulses. This result shows that the exploration of metastable phases can lead to nonvolatile control of physical properties.
We are also investigating how to predict which materials are likely to host metastable phases, with the aim of guiding their exploration through materials design. Model calculations have shown that competition between interactions stabilizing different ordered states can create an activation barrier along the phase-transition pathway, allowing metastable phases to become long-lived even in systems without long-range atomic diffusion. Based on these design principles, we aim to explore metastable metallic and superconducting phases and develop nonvolatile switches that exploit their controlled creation and erasure.