AI-Driven Material Achieves 90% Capture of Toxic Iodine in Nuclear Waste

In a significant advancement for nuclear waste management, researchers from the Korea Advanced Institute of Science and Technology (KAIST) have developed a novel material capable of capturing over 90% of radioactive iodine, specifically isotope I-129, from contaminated water. This breakthrough, made possible through artificial intelligence (AI), could revolutionize the cleanup processes associated with nuclear energy by addressing one of its most persistent and hazardous byproducts.

Radioactive iodine-129, with a half-life of approximately 15.7 million years, is notoriously difficult to remove from the environment. Its mobility poses substantial environmental and health risks, making its effective management a critical concern in the nuclear sector. The KAIST team, led by Professor Ho Jin Ryu from the Department of Nuclear and Quantum Engineering, collaborated with the Korea Research Institute of Chemical Technology (KRICT) to create this innovative solution.

The research, published in the Journal of Hazardous Materials, highlights the use of layered double hydroxides (LDHs) — compounds recognized for their structural flexibility and ability to trap negatively charged particles such as iodate (IO₃⁻), the prevalent form of radioactive iodine in aqueous environments. The team leveraged machine learning to identify optimal combinations of metals within LDHs, significantly streamlining the research process. Rather than manually testing thousands of material combinations, they utilized experimental data from 24 binary and 96 ternary compositions to train an AI model for predicting the most promising candidates from a vast pool of metal combinations.

Through this approach, the researchers discovered a quinary compound composed of copper, chromium, iron, and aluminum, designated as Cu₃(CrFeAl). This material exhibited over 90% efficiency in removing iodate from water, demonstrating a marked improvement over traditional silver-based absorbents, which often struggle to effectively capture iodate. Remarkably, the researchers only needed to test about 16% of all possible material combinations to identify Cu₃(CrFeAl), showcasing the transformative potential of AI in nuclear environmental research.

“This study shows the potential of using artificial intelligence to efficiently identify radioactive decontamination materials from a vast pool of new material candidates,” stated Professor Ryu. “It is expected to accelerate research for developing new materials for nuclear environmental cleanup.”

The KAIST team has filed a domestic patent application for their newly developed powder technology and is currently pursuing an international patent application. They are also actively seeking academic and industrial partnerships to further develop iodine-absorbing powders and water filters for practical use at contaminated nuclear sites.

The implications of this research extend beyond immediate environmental cleanup solutions. As the global community grapples with the challenges of nuclear waste management and the ongoing demand for clean energy, the development of effective, AI-driven materials could play a pivotal role in ensuring the sustainability of nuclear power as an energy source.

Future research will focus on enhancing the stability of the Cu₃(CrFeAl) material under real-world conditions, as well as exploring its scalability for industrial applications. The collaborative nature of this research showcases the importance of interdisciplinary approaches, combining insights from nuclear engineering, materials science, and artificial intelligence, to tackle one of the most pressing issues in contemporary energy discourse.

As nuclear energy continues to be a critical element of the global energy mix, advancements such as these promise a more sustainable future where the environmental impacts of nuclear waste can be effectively managed.