| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Alloys and Compounds for Thermoelectric and Solar Cell Applications XIII
|
| Presentation Title |
Developing Data-Driven Design Rules for Thermoelectric Performance of Zintl Phases |
| Author(s) |
Rinkumoni Chaliha, Michael Toriyama, Prashun Gorai |
| On-Site Speaker (Planned) |
Rinkumoni Chaliha |
| Abstract Scope |
Zintl phases are among the best performing thermoelectric (TE) materials due to their unique chemical bonding and electronic structures. Thermoelectric Zintl phases have so far been discovered through chemical intuition, serendipity, or computationally-guided searches. While empirical guidelines have been proposed to rationally design new Zintl phases with high TE performance, there is a need to develop structure-property design rules that relate fundamental compositional and structural properties to their TE performance. In this study, we use a data-driven approach to develop such design rules by considering a large dataset of >1000 Zintl phases, including reported phases in the Inorganic Crystal Structure Database and newly predicted ones. We calculate the transport properties and TE performance of these Zintl phases using semi-empirical models and first-principles computations. Using machine learning models and feature importance algorithms, we reveal how bonding features and local structural motifs determine the TE performance of Zintl phases. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Energy Conversion and Storage, |