KAIST Leverages AI to Identify Ideal New Material for Radioactive Iodine Removal

Managing radioactive waste remains a major challenge in the continued use of nuclear energy, with radioactive iodine presenting particularly serious environmental and health concerns due to its long half-life (15.7 million years for I-129), high mobility, and toxicity.

A Korean research team has now harnessed artificial intelligence to identify a novel material capable of removing radioactive iodine, marking a significant step forward in nuclear environmental remediation. Plans are underway to commercialize the technology through partnerships between academia and industry, developing products ranging from iodine-absorbing powders to filters for treating contaminated water.

KAIST announced on July 2 that Professor Ho Jin Ryu from the Department of Nuclear and Quantum Engineering, in collaboration with Dr. Juhwan Noh from the Digital Chemistry Research Center at the Korea Research Institute of Chemical Technology (KRICT)—under the National Research Council of Science & Technology (NST)—successfully employed AI techniques to discover an efficient new material for iodine removal.

Recent research indicates that radioactive iodine is typically found in water as iodate (IO₃⁻). However, traditional silver-based adsorbents exhibit weak chemical affinity for iodate, limiting their effectiveness. This underscores the need for newly engineered materials with stronger absorption capabilities.

READ ALSO: Revolutionary Robotic Surgery Combines Dual-System Technologies for Groundbreaking Prostate Procedure

READ ALSO: Nitrogen and Argon Plasma Boosts the Effectiveness of Carbon-Based Supercapacitor Electrodes

READ ALSO: Manchester-based researchers contribute to the discovery of an ancient Egyptian city

Professor Ho Jin Ryu’s research team employed a machine learning-guided experimental approach to pinpoint the most effective iodate adsorbents among a class of compounds known as Layered Double Hydroxides (LDHs), which incorporate multiple metal elements.

The team developed a novel multi-metal LDH—Cu₃(CrFeAl), composed of copper, chromium, iron, and aluminum—that demonstrated outstanding performance, removing more than 90% of iodate. This success was achieved by leveraging AI-driven active learning to efficiently navigate a wide range of material compositions, a task that would be highly challenging using traditional trial-and-error methods.

The research team recognized that LDHs, similar to high-entropy materials, can accommodate a broad range of metal combinations and feature structures well-suited for anion adsorption. However, the sheer number of possible multi-metal combinations makes it extremely difficult to determine the optimal composition using conventional experimental techniques.

To address this challenge, the team turned to machine learning. By starting with experimental data from 24 binary and 96 ternary LDH compositions, they extended their exploration to include quaternary and quinary candidates. This AI-driven approach enabled them to identify the most effective material for iodate removal while testing only 16% of all potential combinations.

Professor Ho Jin Ryu noted, “Our study highlights the power of artificial intelligence in efficiently uncovering effective materials for radioactive decontamination from a vast array of possibilities, paving the way for faster progress in developing advanced solutions for nuclear environmental remediation.”

The research team has submitted a domestic patent application for their newly developed powder technology and is in the process of applying for an international patent. They aim to further improve the material's effectiveness under different environmental conditions and advance toward commercialization by collaborating with industry partners to develop filters for treating radioactive-contaminated water.