We show that an agentic large language model (LLM) (OpenAI o3 with deep research) can autonomously reason, write code, and iteratively refine hypotheses to derive a physically interpretable equation for competitive adsorption on metal-organic layers (MOLs)—an open problem our lab had struggled with for months. In a single 29-min session, o3 formulated the governing equations, generated fitting scripts, diagnosed shortcomings, and produced a compact three-parameter model that quantitatively matches experiments across a dozen carboxylic acids.
This is really impressive! It’s fascinating to see an agentic LLM autonomously reason through a complex physical chemistry problem, generate code, and refine hypotheses—all in under half an hour. The fact that it produced a compact, experimentally validated model for competitive adsorption shows huge potential for combining AI with real-world lab research. Excited to see how this approach develops further!