Coupling High-Throughput Density Functional Theory, Automated Experimentation, and Adaptive Experimental Design To...

Area of Science:

• Materials Science
• Chemical Engineering
• Computational Chemistry

Background:

• Developing novel separation technologies for rare-earth elements is crucial for the critical materials supply chain.
• Current methods often lack the efficiency and selectivity required for complex mixtures.

Purpose of the Study:

• To develop a pH-controlled solvent extraction strategy for selective lanthanide separations.
• To integrate computational, machine learning, and automated experimental workflows for accelerated discovery.

Main Methods:

• Utilized database screening, density functional theory (DFT) calculations, and automated high-throughput experiments.
• Employed multi-objective Bayesian Optimization to explore a multidimensional chemical matrix (pH, extractant, holdback agent, salt concentrations).
• Identified oxaloacetic acid as a holdback agent with di(2-ethylhexyl)phosphoric acid (HDEHP or D2EHPA) extractant.

Main Results:

• Achieved selective extraction of neodymium (Nd), europium (Eu), dysprosium (Dy), and holmium (Ho).
• Demonstrated selective separation of Eu, Dy, and Ho over Nd at pH ~2.0.
• Showcased selective separation of Eu from Dy and Ho at pH ~0.5.
• Increased separation factors by 4-fold compared to HDEHP-only systems using Bayesian Optimization.

Conclusions:

• Established a hierarchical, data-driven framework for identifying selective separation conditions.
• The integrated workflow significantly accelerates separations discovery and design for rare-earth elements.
• This approach provides a foundation for optimizing critical materials separation processes.