Search

Machine learning is becoming a valuable tool for scientific discovery. Particularly attractive is the application of machine learning methods to the field of materials development, which enables innovations by discovering new and better functional materials. To apply machine learning to actual materials development, close collaboration between scientists and machine learning tools is necessary. However, such collaboration has been so far impeded by the black box nature of many machine learning algorithms. It is often difficult for scientists to interpret the data-driven models from the viewpoint of material science and physics. Here, we demonstrate the development of spin-driven thermoelectric materials with anomalous Nernst effect by using an interpretable machine learning method called factorized asymptotic Bayesian inference hierarchical mixture of experts (FAB/HMEs). Based on prior knowledge of material science and physics, we were able to extract from the interpretable machine learning some surprising correlations and new knowledge about spin-driven thermoelectric materials. Guided by this, we carried out an actual material synthesis that led to the identification of a novel spin-driven thermoelectric material. This material shows the largest thermopower to date.

In the previous work, it is reported that the Spin-Seebeck effect (SSE), which refer to the generation of a spin current from a temperature gradient, can be enhanced by Fe interface treatment. Here, we investigated the Fe thickness (dFe) dependency of spin-Seebeck voltage (VSSE) and mixing conductance (gr) in Pt/Fe/Bi:YIG/SGGG system. As a result, magnitude of VSSE had a peak at dFe ⇌ 1 ML (monolayer , ⇌ 0.3 mm), and also increase of gr was saturated at this point. It suggests that VSSE increase with increasing gr when dFe is smaller than 1.0 ML. For the case in which dFe is larger than 1.0ML, however, VSSE decreases due to a spin current decay in Fe layer with a constant gr. These experimental results are consistent with previous theoretical works.

We have developed an infrared image sensor based on bolometers on a plastic substrate that is potentially ultra low-cost due to its simple structure consisting of mainly organic materials. The bolometer includes a thermal isolator made of a very low thermal conductive polymer (parylene) and a newly developed, highly sensitive (TCR

Heat-flow sensing is expected to be an important technological component of smart thermal management in the future. Conventionally, the thermoelectric (TE) conversion technique, which is based on the Seebeck effect, has been used to measure a heat flow by converting the flow into electric voltage. However, for ubiquitous heat-flow visualization, thin and flexible sensors with extremely low thermal resistance are highly desired. Recently, another type of TE effect, the longitudinal spin Seebeck effect (LSSE), has aroused great interest because the LSSE potentially offers favourable features for TE applications such as simple thin-film device structures. Here we demonstrate an LSSE-based flexible TE sheet that is especially suitable for a heat-flow sensing application. This TE sheet contained a Ni0.2Zn0.3Fe2.5O4 film which was formed on a flexible plastic sheet using a spray-coating method known as “ferrite plating”. The experimental results suggest that the ferrite-plated film, which has a columnar crystal structure aligned perpendicular to the film plane, functions as a unique one-dimensional spin-current conductor suitable for bendable LSSE-based sensors. This newly developed thin TE sheet may be attached to differently shaped heat sources without obstructing an innate heat flux, paving the way to versatile heat-flow measurements and management.

Thermoelectric converters based on the spin Seebeck effect (SSE) have attracted great attention due to their potential to offer novel applications such as energy harvesting and heat-flow sensing. For converting a SSE-induced spin current into an electric current, a transition metal film such as Pt, which exhibits large inverse spin-Hall effect (ISHE), has been typically used. In this work, we show an all-oxide SSE device using ruthenium oxide (RuO2) as a conductive film. We found that both the sign and magnitude of the SSE-induced ISHE voltage V appearing in the RuO2 film changes depending on the post annealing temperature, and that the magnitude can become larger than that of a standard SSE device using Pt. The similar sign change was also observed in Hall-resistance measurements of the RuO2 films. X-ray absorption fine structure (XAFS) spectra of as-deposited and annealed RuO2 revealed that the annealing process substantially improved the long-range crystalline order in RuO2. This suggests that change in the crystalline order may modify the dominant ISHE mechanism or electronic states in RuO2, leading to the sign reversal of V as well as the Hall coefficient. Our result demonstrates that RuO2 is an interesting material not only as a practical ISHE film but also as a testbed to study physics of spin-to-charge converters that depend on their crystalline order.

We have developed a high-throughput, two-dimensional-mapping (isoelectric point [pI], mass-to-charge ratio [m/z]) method by combining a capillary isoelectric focusing chip sealed with removable resin tape and a matrix-assisted laser desorption/ionization time-of-flight mass spectrometer. Sample proteins are separated in a meandering channel on the chip and immediately frozen. The tape is then removed and the proteins are freeze-dried. The freeze-drying maintains the separation state of the proteins and prevents movement of the sample solution, which can reduce pI resolution. A matrix solution is then applied and mass spectrometry is carried out by laser irradiation. The whole process takes less than 70 min, more than 10 times faster than with two-dimensional, polyacrylamide gel electrophoresis.

Thermoelectric conversion (TEC) technologies, which convert heat into electricity, have received a great attention, because they are expected to be a powerful approach to utilize wasted thermal energy. Here we present novel thermoelectric converters based on the spin Seebeck effect (SSE), and show their scaling law which is largely different from that of conventional TEC devices. We experimentally demonstrate that the TEC output signals straightforwardly increase with the size of the converters. This scaling law enables us to implement simple-structured thermoelectric converters by using productive film-coating methods. Such coating-based TEC techniques may pave the way for a wide range of applications using a variety of heat sources.

We have developed an infrared sensor for low-cost infrared imaging devices based on bolometers on a plastic substrate. The bolometer includes a thermal isolator made of polymer, and a newly developed high-sensitive (TCR > 2 %/K) carbon nanotube thin film thermistor. This structure can be fabricated with print processes, and additionally, enables us to avoid the conventional high-cost vacuum packaging process. The infrared irradiation test showed that the responsivity of 340 V/W was obtained.

This paper describes an electrochemical immunosensor for the detection of target molecules without the use of labeled antibodies. The sensor is based on the combination of a gold electrode covered with a conductive self-assembled monolayer (SAM) of mixed thiols and a permeability measurement of an antibody layer anchored on the SAM. The permeability was measured by the square wave voltammetry (SWV) of ferrocyanide added in a measurement solution. The conductive SAM allows constant electron transfer between the electrode and ferrocyanide, providing a stable signal during repeated SWV measurements in contrast to immunosensors using conventional SAM techniques. We demonstrate that our sensor using the Anti-human chorionic gonadotropin (hCG) antibody selectively detects hCG in a minute without any washing process. The presented label-free immunosensor technology is fundamentally suited for easy-to-use assays for biological molecules such as hormones and drugs.

A size-based pre-fractionation chip applicable for enrichment of macromolecules in biological samples is proposed. In this chip, a fractionation unit with a nano-pillar anisotropic array, which functions as sieve meshes, is installed at a bi-forked junction. The mesh sizes are close to the diameters of macromolecules and vary according to the direction to the main and branch exit channels. This fractionation unit provides continuous-flow fractionation without both the dilution of samples and the limitations with the sample loading volume. Additionally, the small channel-volume, which is characteristic of chip-based systems, can reduce sample loss. Using nanometer-scale fabrication technology, we fabricated the chips and demonstrated the flow direction control of DNA molecules.