Towards application of materials informatics to development of thermal functional materials.
Professor Junichiro Shiomi et al. from The University of Tokyo aimed to reduce the thermal conductivity of semiconductor materials by reducing the internal nanostructure, and successfully minimized thermal conductivity by designing, fabricating, and evaluating the optimal nanostructure-multilayer materials through materials informatics (MI), which combines machine learning and molecular simulation.
In 2017, this research group developed a method to design an optimal structure that minimizes or maximizes thermal conductivity via MI based on computational science. However, it has not been experimentally demonstrated, and preparation of nano-scale structures and realization of an optimal structure based on property measurements were desired.
Thus, the research group utilized a film deposition method able to regulate, at a molecular level, a superlattice structure wherein two materials were alternately layered at several nm thick, and a measurement method that could assess thermal conductivity of a film at nano-scale, and realized the optimal aperiodic superlattice structure that minimizes thermal conductivity. With the optimal structure, wave interference of the lattice vibration (phonon) that conducts heat was maximized, and thermal conductivity was strongly regulated.
To read the full article, click here.