A team of researchers from Oak Ridge National Laboratory, Cleveland Clinic, and IBM has achieved the first-known quantum computation of a fusion fuel material, a molten salt called FLiBe that future nuclear fusion reactors will use to breed their own fuel.
The work, published as an arXiv preprint, used a hybrid quantum-classical approach called quantum-centric supercomputing to calculate nine molecular configurations of FLiBe (a salt of fluorine, lithium, and beryllium). Quantum circuits handled the electron behavior, the bonding and interactions between atoms, which is ideally suited to quantum hardware, while classical computers completed the remaining calculations.
Why FLiBe matters
One of the biggest engineering challenges facing commercial fusion energy is tritium scarcity. Tritium, a radioactive isotope of hydrogen, is a key fusion fuel but is extremely rare in nature. Future fusion reactors are designed to breed their own tritium inside a “blanket” of molten salt surrounding the reactor core. FLiBe is the leading candidate for that blanket material.
Understanding exactly how tritium atoms bind with FLiBe at the atomic level is essential for optimizing reactor design and tritium production rates. Until now, these interactions could only be approximated with classical computing methods.
“The work builds on our advances in simulating complex biological systems at scale, including proteins spanning 12,635 atoms, and extends those techniques into materials science,” said Kenneth Merz of Cleveland Clinic, a corresponding author on the paper.
The technical approach
The team used IBM’s quantum-centric supercomputing architecture, which combines quantum processors for the parts of the calculation that benefit from quantum advantage, electron correlation and bonding, with classical supercomputers for the rest. The same methodology was previously applied to protein simulations and is now being adapted for materials science.
Jerry Chow, IBM’s CTO of Quantum-Centric Supercomputing, said the work demonstrates that “bringing quantum, AI, and classical computing together is essential to tackling our society’s most fundamental scientific challenges.”
The next steps include reducing data transfer time between quantum and classical processors and expanding the size of molecular systems that can be modeled. The researchers aim to make the workflow available to fusion developers for their own reactor materials.
Sources: Interesting Engineering (July 8); arXiv (2026)