Towards AI-Driven Recommendation of Liquid Chromatography Conditions for Chemical Reactions

Youngchun Kwon, Hyukju Kwon, Jinju Park, Seokho Kang, Youn-Suk Choi

Additional Submission Instructions: For the camera-ready version, please include the author names and affiliations, funding disclosures, and acknowledgements.

Track: Track 1: Original Research/Position/Education/Attention Track

Keywords: Liquid Chromatography, AI Agent, Chemical Reaction Analysis

TL;DR: An LLM-based multi-agent system recommends liquid chromatography conditions to detect and separate diverse components in a complex chemical reaction within a single run.

Abstract: Liquid Chromatography (LC) is a foundational tool for chemical analysis, but applying it to complex reactions is challenging. The primary difficulty lies in establishing conditions that achieve simultaneous detection and chromatographic resolution for multiple reactants and products with diverse properties. In this work, we investigate the capability of Large Language Models (LLMs) in agentic LC condition recommendation for comprehensive chemical reaction analysis. We present an LLM-based multi-agent system comprising chemistry-aware sub-agents that emulate the decision-making process of an analytical chemist. The system processes a chemical reaction and user-defined analytical requirements expressed in natural language to autonomously search relevant literature, reason over compound properties, and propose plausible LC conditions that can detect all reaction components within a single analytical run. We demonstrate its effectiveness through a case study on organic electronic materials. The source code is available at https://github.com/seokhokang/lc_agent.

Submission Number: 6