JiongYang*

1Materials Genome Institute, Shanghai University, Shanghai, 200444, China;

2Zhejiang Laboratory, Hangzhou, Zhejiang 311100, China

EXTENDED ABSTRACT: Since 2011, there have been emerging several computational material databases, including the Materials Project, AFLOW, Atomly, and MatHub-3d(2d) from MGI, Shanghai University. MatHub-3d aims at the electrical and thermal transport properties; the repository contains over 80,000 inorganic structures, over 30,000 electronic structures, and over 10k electrical transport properties.[!] This talk will cover several recent high-throughput screening works for new thennoelectrics based on the MatHub-3d. In the search of new thennoelectrics, besides the screening from direct calculations of transport properties, the search can also be accelerated by the key features. Based on the previous study that the carrier type of semiconductors relate to the band edge energy position, we propose the strategy of searching n-type Cu-contained compounds by looking for compounds with low-energy conduction band minima. Follow-up theoretical and experimental studies confirm this strategy, and obtain a new thermoelectric compound Culn5Se8 with experimentally ZT of 0.84.[2] In another study, we propose a new concept of "Functional-unit design", borrowed from the "fragment design" in the search of new drug. A linear triatomic bonding is found to be an effective thermoelectric functional unit. This unit is theoretically proved to cause large lattice anhannonicity and band anisotropy, both favoring the the thmoelectric performance. The predicted compound K5CuSb2 has a ZT value of 1.3.[3] The two case studies demonstrate the powerfulness of the combination of material repository and key feature search in accelerating the high-throughput works.

Keywords: MatHub-3d; key feature; energy position; functional-unit design