[15p-K307-1~14] 9.4 Thermoelectric conversion

5:00 PM - 5:15 PM JST(8:00 AM - 8:15 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[15p-K307-13] Exploring high-performance Al-O-based thermoelectric materials working at high temperatures using first-principles calculations in combination with machine learning

〇Rongrong Bi¹, Jun Onoe¹, Toshiaki Nishii², Yusuke Noda³ (1.Nagoya Univ., 2.J-POWER, 3.Kyushu Inst. Technol.)

Keywords:

high-performance thermoelectric aluminum oxides,first-principles calculation,machine learning

A first-principles calculation combined with machine learning was used to search for high-performance Al-O-based thermoelectric materials capable of high-temperture operation at 900K. Among the 3,447 Al-O-based materials registered in the Materials Project, the p-type ZT was calculated for 109 materials for which bulk modulus values were published. These were used as training data to optimize the neural network model, and the ZT of the remaining 3,338 materials was predicted, which results in the efficient extraction of seven candidates whith high ZT_cal.