Augmenting Large Language Models for Automated Discovery of F-Element Extractants

Efficient separation of f-elements is a critical challenge for a wide range of emerging technologies. The chemical similarity among these elements makes the development of selective solvent extraction reagents both slow and difficult. Here, we present a quasi-autonomous AI-enabled workflow for the design and computational screening of selective extractant ligands. Molecular design is guided by SAFE-MolGen, a large language model-based agentic system that leverages curated extraction data to propose new ligands and preliminarily rank their performance using a supervised machine learning model trained on experimental data sets to consider the impact of realistic experimental conditions. Promising human-approved ligands are then passed to a second automated pipeline that constructs three-dimensional metal–ligand complexes and performs quantum mechanical free energy calculations to directly assess the metal selectivity. We demonstrate this approach for Am(III)/Eu(III) separations and report several newly designed ligands predicted to exhibit higher Am(III)/Eu(III) selectivity than the benchmark extractant CyMe$_4$BTBP. This workflow accelerates computational exploration of the molecular space in this data-sparse field and provides a general strategy for the rapid generation and evaluation of novel lanthanide (Ln) and actinide (An) extractants.