High performance computing: Advances and challenges in modeling rare Earth elements and actinides in 2023

Computational science applications to model rare earth elements (REEs), lanthanides, and actinides are essential to address needs in national and nuclear security. Computationally-driven findings are critical in many applications, including to advance radiotherapeutics design, optimization of separations in REE production and purification technologies, and radiological waste management.

High performance computing (HPC) advances in recent decades have enabled a multitude of solutions to challenges in global needs, including weather predictions, green energy, medical therapeutics, and materials science. A large contribution has emerged from applications of artificial intelligence (AI) – including machine learning (ML) and deep learning. Having entered the exascale era of computing, software technologies to model REEs and actinides face challenges due to an imbalanced hardware-software ecosystem that presents limitations to obtain accurate predictions of REE- and actinide-containing compounds at large scale. Challenges in scalability, performance, and memory limitations restrict efficient modeling in nuclear- and radiochemical applications.

This presentation will address a historical perspective of HPC resources development since the 1990s, current tools to model REEs and actinides, computational protocols for preferential binding in radiochemical separations, and applications of AI/ML in binding selectivity. Additionally, current contributions to exascale efforts in multidisciplinary work involving co-design will be highlighted.