Thermoelectrics is the direct conversion of heat (or a temperature gradient) to electricity and vice versa. It exploits two related mechanisms: the Seebeck effect and the Peltier effect (Fig. 1).


Fig. 1 Schematic diagram for thermoelectric

Thermoelectric devices are solid-state energy converters that typically consist of many pairs of n- and p-type semiconductors connected electrically in series and thermally in parallel (Fig 2).


Fig. 2 Typical thermoelectric device consisting of p- and n-type semiconductor (green and red represent opposite type) pairs

The energy conversion efficiency of a thermoelectric device largely depends on that of the materials used, so expansion of the practical usage of the thermoelectric effect requires highly-efficient materials.

Materials based on Bi2Te3 and PbTe have high efficiency and have been commercialized. However, Pb and Te are toxic and expensive. Thus, we focus on the development and design of relatively safe and inexpensive and materials.

To this end, we are developing a new class of thermoelectric materials by using artificial intelligence (AI). The Algorithm was trained on existing experimental data with thermodynamic parameters extracted from their composition to classify solid-state materials (Fig. 3). AI is expected to predict the phase of materials in random combinations, and to be a powerful tool for materials design.


Fig. 3 Schematic flow of design for new class thermoelectric materials assisted by AI