Energy Conversion Technologies
MANA Scientists Use Machine Learning for Thermoelectric Materials
Debra John — January 17, 2025Researchers at the Research Center for Materials Nanoarchitectonics (MANA) have used machine learning to improve the discovery of kesterite-type thermoelectric materials, which convert waste heat into electricity. Made from non-toxic, widely available elements, these materials, such as "Cu₂ZnSnS₄, perform well at temperatures between 300 and 800K (26 to 526°C)." However, finding optimal manufacturing conditions is time-consuming.
To accelerate this process, MANA researchers applied machine learning to optimize sintering. In four experimental cycles, they "improved the thermoelectric performance of Cu₂.₁₂₅Zn₀.₈₇₅SnS₄ by 60% and was led by Dr. Cedric Bourges, the study involved multiple researchers from MANA and NIMS, Japan — using data from 11 spark plasma sintered samples, they achieved a record thermoelectric efficiency (zT) of 0.44 at 725K."
This approach may enhance other materials used in solar panels, batteries, and electronics, contributing to advancements in energy conversion technology.
Image Credit: Research Center for Materials Nanoarchitectonics (MANA) / National Institute for Materials Science (NIMS)