Researchers from the University of Tokyo’s Institute of Industrial Science and collaborators have proposed “flip engineering” that uses machine learning to design the atomic-layer stacking configurations of Janus MoSSe, improving thermal management between layers in 3D chips. The 14-layer stacking includes 8,192 possible flip sequences. After evaluating about 3% of the configurations, the team identified a structure with an in-plane thermal conductivity approximately 760 times its effective interlayer thermal conductivity at room temperature—a ratio about 35% higher than that of an ordered crystal stacking of the same thickness. The study was published in ACS Nano, with manufacturing and experimental validation identified as subsequent challenges.
Source:nanowerk.com · Read the original article
Leave a Reply