Monday , May 20 2019
Home / canada / Tech: Choosing an optimal structure of about 8 billion candidates may make the company more energy efficient – (report)

Tech: Choosing an optimal structure of about 8 billion candidates may make the company more energy efficient – (report)



Scientists have designed a multi-layer metamaterial that understands selectively selective wavelength emission spectra by combining the learning of the machine (Bayesian optimization) and thermal emission properties calculations (electromagnetic calculation). The team then co-fabricated the metamaterial experiment designed and intimidated the performance. These results may facilitate the development of efficient energy devices.

NIMS, Tokyo University, Niigata University and RIKEN have jointly designed a multilayered metamaterial that understands ultra-narrowband selective wavelength and thermal emission by combining a machine learning (BASIC optimization) and thermal emission characteristics calculations (electromagnetic calculation). The team then co-fabricated the metamaterial experiment designed and intimidated the performance. These results may facilitate the development of efficient energy devices.

Thermal radiation, a phenomenon that emits heat object as electromagnetic waves, is possible to apply to a variety of energy devices, such as selective wavelength furnaces, infrared sensors and thermophotovoltaic generators. High efficient thermal emiters should display emission spectra with narrow bands in an almost usable wavelength wave range. The development of such efficient thermal emitters has been targeted by numerous studies using metamaterials that can manipulate electromagnetic waves. However, most of them took an approach of characterizing material structures selected empirically. , It is difficult to identify the optimal structure from a large number of candidates.

The joint research group has developed a method for designing metamaterial structures with optimal performance of thermal radiation through a combination of computerized learning and calculation of thermal emission properties. This project focused on easy-fabricated multilayered metamaterial structures consisting of three types of materials and 18 layers of varying thickness. The application of this method to eight billion candidate structures led to the prediction that a nanostructure consisting of not periodically arranged semiconductor and dielectric materials would perform excellent thermal radiation, which was contrary to conventional knowledge. Then the research group actually fabricated the metamaterial structure and measured its thermal emission spectrum, resulting in the band demonstrating the narrowest thermal emission. Measured in terms of Q factor (parameter used to measure the width of thermal emission spectral bands), the newly designed nanostructure produced a Q-factor close to 200, where 100 were considered the upper limit for conventional materials? Northern band thermal emission.

This study has demonstrated the effectiveness of learning in machines developing efficient thermal emission metamaterials. The development of metamaterials with desirable thermal emission spectrum is expected to allow for more efficient use of energy throughout the company. Because of the nanostructure design method developed and applicable to all kinds of materials, it may be used as an effective tool for the planning of high-performance materials in the future.

source:

National Institute of Materials Science, Japan. .


Source link