| About this Abstract |
| Meeting |
2020 TMS Annual Meeting & Exhibition
|
| Symposium
|
Thermal Transport in Crystalline and Non-crystalline Solids: Theory and Experiments
|
| Presentation Title |
Physics-guided Machine-Learning Design of Aperiodic Superlattices with Maximum Localization of Coherent Phonons |
| Author(s) |
Pranay Chakraborty, Tengfei Ma, Yan Wang, Lei Cao |
| On-Site Speaker (Planned) |
Pranay Chakraborty |
| Abstract Scope |
Aperiodic superlattices exhibit much lower lattice thermal conductivity than their periodic counterparts due to the localization of coherent phonons. However, finding the optimal configuration, i.e., layer thickness distribution and order of the thicknesses, to achieve the lowest possible thermal conductivity has been a daunting task. This primarily arises from a lack of knowledge of how superlattice configuration affects the behavior of coherent phonon transport and localization. In this work, we have identified several structural parameters that are strongly correlated with the lattice thermal conductivity of the aperiodic superlattice using classical molecular dynamics simulations and atomistic Green’s function simulations. We have revealed that they affect the coherent phonon band structure and thus transmission significantly. Moreover, we have found that using physics-guided machine learning, which considers both configuration and the structural parameters identified through this work altogether, can predict the thermal conductivity of aperiodic superlattice more accurately and efficiently. |
| Proceedings Inclusion? |
Planned: Supplemental Proceedings volume |