Hard-to-synthesize materials revived using AI: An LLM-based materials redesign technology

February 2, 2026 - 06:52

A pioneering artificial intelligence system is now capable of redesigning complex, hard-to-synthesize materials, potentially unlocking a new wave of functional compounds for electronics, energy storage, and beyond. Developed by a team led by Professor Yousung Jung at Seoul National University, the technology leverages large language models (LLMs) to intelligently modify molecular structures.

The core challenge addressed is that many theoretically promising materials are rendered impractical due to extreme synthesis conditions, such as requiring intense heat or pressure. The new AI tool navigates this by analyzing a material's fundamental composition and property goals. It then strategically substitutes specific elements within the structure with chemically similar alternatives that are easier to combine and process.

This approach has already proven successful in the lab. Researchers tasked the AI with finding a viable alternative to a specific manganese-based oxide crucial for battery cathodes. The system proposed a new, related compound. Subsequent real-world synthesis confirmed the AI's design was not only easier to produce but also maintained the desired electrochemical performance.

Experts highlight that this method moves beyond simple prediction, entering the realm of intelligent, goal-oriented design. By overcoming the synthesis barrier, the technology could dramatically accelerate the discovery and practical application of advanced materials, shortening development timelines from years to months for next-generation technologies.