Human-Machine Collaborative Design of SnTe-Based Thermoelectric Materials via a Multiagent Framework Leveraging Large...

Area of Science:

• Materials Science
• Artificial Intelligence
• Energy Science

Background:

• Traditional thermoelectric material design relies heavily on empirical methods and computational simulations.
• Developing high-performance thermoelectric materials for energy applications is challenging.
• Existing approaches often lack efficiency in extracting knowledge from vast scientific literature.

Purpose of the Study:

• To develop an innovative framework for human-machine collaborative design of SnTe-based thermoelectric materials.
• To leverage large language models (LLMs) for efficient knowledge extraction and optimization strategy deduction.
• To demonstrate the efficacy of LLM-guided design in enhancing thermoelectric material performance.

Main Methods:

• Implementation of a multiagent collaborative framework driven by LLMs.
• Integration of strategy planning and reasoning modules for knowledge extraction from literature and experimental data.
• Guided experimental synthesis and characterization of SnTe-based materials with Sb, Ge, and Cu incorporation.

Main Results:

• The LLM framework successfully deduced optimization strategies and compositional ranges.
• Incorporation of Sb, Ge, and Cu elements modulated carrier concentration and induced multiscale defects.
• A novel sample achieved a thermoelectric figure of merit (zT) of ~1.2, a 40% increase over Sb-doped SnTe.

Conclusions:

• LLM-driven collaborative design offers a novel paradigm for material discovery, surpassing traditional methods.
• The developed framework significantly enhances thermoelectric material performance.
• This approach has broad implications for high-performance material design in energy fields, including waste heat recovery and solid-state cooling.