Automated Material Discovery With LC–MS

Hyukju Kwon from SAIT, Samsung Electronics, Korea, presented a talk at HPLC 2023 called “The Role of LC–MS for Autonomous Material Discovery” in a keynote lecture at the 51st International Symposium on High Performance Liquid Phase Separations and Related Techniques (HPLC 2023) in Düsseldorf this week (1).

Side view of a humanoid head with blue and yellow eyes and vibrant neon neural network, representing futuristic technology and artificial intelligence. Generative AI | Image Credit: © Sascha - stock.adobe.com

In material development, the concept of autonomous synthesis aims to discover new materials. This previously involved experiment planning, sample preparation, synthesis, purification, and analysis—tasks that can possibly be handled by AI robots.

To go beyond simple system automation and optimize synthetic conditions, it is essential to incorporate a rapid and accurate analysis module that provides real-time data feedback to enhance AI-based platforms. Kwon proposed the integration of liquid chromatography–mass spectrometry (LC–MS) in a batch-type autonomous synthetic platform. To identify the optimal synthetic conditions, one must conduct qualitative and quantitative analyses of the reaction solution and evaluate the necessary reaction kinetics.

Kwon demonstrated that small samples of the reaction solution are extracted at specific time intervals from different reaction vials with varying recipes. These samples undergo dilution, mixing, filtration, and injection into the LC–MS system through a sample preparation device. Once the synthesis of the desired target compound is confirmed by the MS spectrum, the reaction kinetics are calculated based on the conversion ratio of the starting material to the target product. All the results obtained from the LC–MS system are fed back to the AI system to explore better synthetic conditions.

To accommodate diverse analysis conditions based on the reaction scheme, the system has been enhanced to handle up to eight mobile phase solvents and six columns. Programmed modes, including reverse phase, normal phase, size exclusion, and ion exchange, are implemented, with pre-conditioning between each separation mechanism change to ensure system stability. By implementing this autonomous synthetic platform, a well-known synthesis reaction was successfully verified, yielding a higher conversion yield compared to the reference synthesized by human researchers in a short period of time. Importantly, the entire operation and data analysis were conducted without human intervention, highlighting the potential of AI-driven advancements in material development.

Reference

(1) Kwon, H. The Role of LC–MS for Autonomous Material Discovery. Presented at: HPLC 2023. June 18–22, 2023. Duesseldorf, Germany. KN16.