The Atomic Challenge: How AI and Quantum Chemistry Could Tame Nuclear Waste
Russian researchers have combined quantum chemistry and artificial intelligence to identify stable technetium carbide structures that could enable long-term storage of technetium-99, one of the most problematic radioactive isotopes. The work points to a new, predictive approach to nuclear waste management.
A Quantum Breakthrough in Materials Science
Nuclear power remains one of the key sources of low-carbon energy for the future, but its expansion is constrained by a persistent environmental dilemma: how to manage long-lived radioactive waste. One of the most troubling isotopes is technetium-99, a uranium fission product with a half-life of about 211,000 years. The isotope migrates easily through soil and groundwater, posing a long-term threat to ecosystems over geological timescales. Russian scientists now argue they have found a path toward addressing this global challenge, using tools at the intersection of quantum chemistry and artificial intelligence.
A research team from leading Russian institutions – AIRI, Skoltech, RKhTU, and IFKhE RAS, with support from Sberbank – has made a breakthrough in the search for materials capable of safely immobilizing technetium-99. Instead of relying on trial-and-error experiments that can take decades, the scientists applied a neural network model trained on high-precision quantum-chemical calculations. The result is a detailed “stability map” of technetium carbides – effectively an atlas of atomic configurations, including rare structures that classical methods are unable to detect. This allows experimental researchers to deliberately synthesize the most stable compounds, capable of locking the radionuclide in place for geological timescales.
AI as an Ally of Nuclear Safety
The core innovation lies in the synergy between two advanced technologies. Quantum chemistry provides fundamental accuracy in calculating atomic bond energies, while artificial intelligence accelerates the exploration of possible configurations, revealing patterns invisible to the human eye. This approach makes it possible not only to rediscover known phases, but also to predict exotic, thermodynamically favorable configurations.
That capability is especially valuable for technetium, an element that is virtually absent in nature and whose chemistry remains poorly understood. With the help of AI, researchers now have a systematic material-design tool, replacing random searches with targeted atomic-level engineering.
From the Lab to Industry
The practical significance of the work extends beyond fundamental science. Stable technetium carbides could serve as the basis for composite host materials that not only store technetium-99, but also enable its transformation into stable ruthenium-100 through neutron irradiation, effectively “burning off” residual radioactivity.
The methodology developed in Russia can already be applied to other problematic nuclear waste isotopes, such as iodine-129. The interdisciplinary collaboration – spanning academic institutes and a major technology player in the form of Sber – also highlights the maturity of Russia’s ecosystem for translating scientific advances into industrial solutions. Over the longer term, this opens up export potential: countries with advanced nuclear sectors, from France to China, are actively seeking safer waste-management technologies.
The study is more than a single scientific publication. It signals a broader shift in materials science toward a systematic, predictive, and digital discipline. Moving away from empiricism toward AI-optimized material design is likely to become a defining trend of the next decade. Judging by recent results, Russia intends to be among the leaders in this transition. Nuclear energy can only become truly sustainable once its most dangerous byproducts are managed responsibly – and the combination of artificial intelligence and quantum theory is increasingly pointing the way forward.
