Nanoinformatics

We aim to explore the frontier of nanomaterials science and utilize nanostructure information. However, the many challenges are difficult for one group to overcome alone. Hence, we established a target specific materials development (called the common subject or CS) that involves members from different fields. All members are aware of CS tasks. Specifically, there are three CS tasks: (1) Functional ceramic materials, (2) Solid state ionic materials, and (3) Catalytic materials. Under clear motivation, principal investigators, who are active experts in each field, conduct coordinated experiments, theoretical calculations, and analyses using information science techniques. In addition to normal processes, experts in high-pressure/high-temperature processes and atomic layer control also contribute as principal investigators.

These collaborative research projects have contributed not only to the development of new materials but also to the establishment of scientific theories necessary for new material development.

In a supported metal catalyst, the d-band center of the metal is a good indicator to describe the molecular adsorption energy and the rank order of the catalytic activity. However, experimental investigations are difficult, and the value is often obtained by theoretical calculations (DFT calculations). From the viewpoint of high-speed exploration of highly active catalyst materials, computing the d-band center exhaustively for various types of metal and alloy systems is an inefficient method due to the burdensome calculation and time costs.

In this project, we have constructed a regression model of the d-band center values employing machine learning methods and initial data for 121 types of metals and alloys obtained by DFT calculations by Nørskov et al. Readily available explanatory variable (descriptor) parameters such as the group in the periodic table, density, and ionization energy of each metal have been used. We have demonstrated that the prediction can be made much faster without sacrificing the precision compared to the situation where all individual values are obtained by DFT calculations.

This technique should predict the d-band center of various other metal/alloy systems, allowing candidate materials to be rapidly identified. In addition, information on activity ranking and maximum activity can be found.

I. Takigawa et al. RSC. Adv. 6 (2016) 52587.

A protective film that shields the surface of a structural member is indispensable to prolong the life in a high-temperature corrosive environment. For example, polycrystalline alumina is used as a protective film for heat resistant members exposed to combustion gas environments such as aircraft engines. To further improve a protective film’s performance, it is extremely important to understand the essence of mass transfer through grain boundaries (high-speed diffusion paths) in a film and precisely control the movement.

In this field, we perform high-temperature oxygen permeation tests using polycrystalline wafers or twin crystal wafers cut from alumina-sintered bodies as model films. Then we evaluate and analyze the mass transfer mechanism in the film under an oxygen potential gradient (dµO). In addition, we evaluate and analyze the electronic state of twin grain boundaries and defect formation energies using first-principles calculations and electronic energy loss spectroscopy (EELS) to clarify the correlation between the grain boundary structure and mass transfer experimentally and theoretically.

We have identified that semiconductive characteristics increase at grain boundarieswith a low conformity such as polycrystalline grain boundaries. The change in the electronic state at such a grain boundary greatly influences oxygen diffusion. We also discovered that the grain boundary diffusion coefficient of oxygen in the film exposed to high temperature dµO drops to about 1/10 compared to the case of self-diffusion (without dµO). This is attributed to an increased electron conductivity contribution of the film. Furthermore, utilizing these characteristics, the mass transfer in laminated films composed of multiple oxides with greatly different electric conduction characteristics can be effectively controlled.

This research aims to improve the performance of protective film, including alumina.

T. Ogawa et al. Acta Mater. 69 (2014) 365.

Materials in which ions diffuse at a high speed in solids are called ionic conductors, also known as solid state electrolytes. They have potential in numerous electrochemical devices such as lithium batteries and fuel cells. Development of a new ionic conductor may realize devices that utilize new electrochemical reactions.

In this project, we focus on hydride ions, which are anions of hydrogen. They are found in La-Sr-Li-H-O–based oxyhydrides, and only hydride ions diffuse in the solid. Pure hydride ion conductivity in oxyhydrides has yet to be reported. Hence, we are exploring a new field of solid chemistry. This new material has a K₂NiF₄-based crystal structure, and controlling its composition improves the ionic conductivity. We have fabricated an all-solid-state device consisting of Ti/La-Sr-Li-H-O–based electrolyte/TiH₂. As a result of discharging at 300℃, we have confirmed the potential response and structural change corresponding to the (de-)hydrogenation reaction of a Ti electrode. Additionally, we have successfully operated an electrochemical device using hydride ions as charge carriers.

Since the oxidation-reduction potential of a hydride ion is −2.25 V (vs. SHE), promoting the development of electrodes and catalyzers should realize a high–energy-density electrochemical device in the future.

G. Kobayashi et al. Science 351 (2016) 1314.

Thermoelectric conversion has attracted attention as a technology to recycle waste heat. However, current thermoelectric materials are expensive, have poor thermal and chemical stabilities, and are sometimes toxic. Hence, they are not applicable to large-scale practical applications. By contrast, metal oxides are attractive thermoelectric material candidates, but one obstacle is their very poor thermoelectric properties.

In this project, we aim to greatly improve their thermoelectric properties by increasing the two-dimensional nature of oxide two-dimensional electron system artificial superlattices. First, our experimental and theoretical investigations on the thermoelectric property of SrTiO₃–SrNbO₃ solid solution found that the de Broglie wavelength greatly differs at the point of Ti/Nb = 1/2. In addition, band calculations revealed that the two-dimensional property is maintained even in the [1 uc SrNbO₃|10 uc SrTiO₃] artificial superlattice. By actually preparing many artificial superlattices via pulsed laser deposition method and investigating their thermoelectric properties, we achieved an output factor of 5 mW m⁻¹ K⁻², which corresponds to double the value of the bulk.

This is the first result to clearly demonstrate that increasing the two-dimensional nature effectively improves the performance of thermoelectric materials.

Y. Zhang et al. J. Appl. Phys. 121 (2017)185102.