.png){kind=icon}
.png){kind=icon}
Return
Data-driven discovery of high-performance multicomponent solid solution thermoelectric materials
.png){kind=icon}
PRE
AI
DOI：10.1016/j.mtener.2022.101070.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The discovery and exploration of novel thermoelectric materials over the past decades have relied primarily on the inefficient Edisonian trial and error approach. It is urgent to develop intelligent approaches like machine learning or data-driven screening to accelerate the process of discovery and optimization of high-performance thermoelectric materials. In this study, by taking the Cu-2(S, Se, Te) solid solutions as an example, we demonstrate a successful data-driven screening approach to reveal the optimal composition ranges with high thermoelectric performance. Based on the known experimental data of Cu-2(S, Se, Te) and their solid solutions, we predicted the contouring diagrams of crystal structures and thermoelectric properties of Cu-2(S, Se, Te) multicomponent materials. Following the prediction, we fabricated a series of Cu-2(S, Se, Te) quaternary solid solutions and systematically investigated their crystal structures, phase transitions, and especially thermoelectric properties. A peak zT of 1.3 at 1000 K is achieved in Cu2S0.4Se0.3Te0.3, which is well in the predicted optimal composition range. We expect this simple yet efficient strategy to be widely applied to the quick screening of other high-performance thermoelectric materials. (c) 2022 Elsevier Ltd. All rights reserved.
Keywords:
Electricalproperties
Thermalconductivity
Copperchalcogenide
Liquid-like
Optimalcomposition
Journal
IF:
8.6
Papers:
2.3K
Citations:
1.2W