Natural-Language-Interfaced Robotic Synthesis for AI-Copilot-Assisted Exploration of Inorganic Materials

Area of Science:

• Materials Science
• Chemistry
• Robotics
• Artificial Intelligence

Background:

• Traditional chemical synthesis automation faces limitations in flexibility and requires specialized expertise.
• Closed-loop systems often employ rigid workflows, hindering broader applicability and rapid discovery.

Purpose of the Study:

• To develop a flexible and adaptive chemical synthesis platform using artificial intelligence (AI) and robotics.
• To streamline the process from conceptualization to experimental execution in materials discovery.

Main Methods:

• Integration of a large language model (LLM) to translate natural language descriptions into executable robotic operations.
• Incorporation of AI-driven literature searches, real-time experimental design, and conversational human-AI interaction.
• Utilizing a modular robotic platform for diverse unit operations such as temperature control, liquid handling, and filtration.

Main Results:

• Successful synthesis of 13 compounds across four distinct inorganic material classes: coordination complexes, metal-organic frameworks, nanoparticles, and polyoxometalates.
• Discovery of a novel family of manganese-tungsten (Mn-W) polyoxometalate clusters.
• Demonstration of AI-enhanced robotics for adaptable and generalizable materials innovation.

Conclusions:

• AI-integrated robotic explorers offer a flexible and adaptive approach to chemical synthesis, accelerating materials discovery.
• This platform simplifies complex experimental workflows, making advanced synthesis more accessible.
• The system shows significant potential for advancing materials innovation through automated, AI-driven experimentation.