HWNEWS
A Berkeley Lab team found that the host metal, not extra heat, drives a low energy fusion boost. It hints at compact neutron sources for medicine and cargo screening.
In a benchtop experiment at Lawrence Berkeley National Laboratory, a thin foil of palladium or titanium hosted a fusion reaction that ran 18 orders of magnitude stronger than the same reaction should have, with no extra heat to explain it. The result, published July 18 in Nature Communications, is not a step toward fusion power. It is something stranger: evidence that the host material is part of the reaction.
The team, led by Arun Persaud of Berkeley Lab's ATAP Division, the group that builds particle accelerators and related physics instruments, packed deuterium, a heavy form of hydrogen, into thin foils of palladium and titanium. They fired a low-energy deuterium-ion beam at the loaded foils and counted the fusion events: deuterium-deuterium reactions that spit out helium-3 and a neutron. Standard nuclear theory says that as the beam energy drops, the fusion rate should plummet on a steep exponential curve, because the positively charged deuterium nuclei have to overcome their mutual electrical repulsion. Below about 2.5 kiloelectronvolts, a unit of beam energy where the fuel barely reacts, that suppression is supposed to be punishing.
In some of the foils, it was not. The yield flattened into a plateau and, at the lowest energies measured, ran about 10^18 times higher than the bare-nucleus prediction. A factor of 10^18 is a 1 followed by 18 zeros, the kind of multiplier that separates the number of grains of sand on a beach from the number of stars in the observable universe. The same reaction, in the same physics, with no material around it, would have produced a vanishingly small signal.
The team describes the result as a new knob. "For a long time, we thought of the host material as something the fuel sits inside," Persaud said, according to the Berkeley Lab release. "What we are learning is that the host material can also act like a catalyst, giving us a new knob to turn." The chemistry analogy is deliberate. A catalyst changes how fast a reaction runs without being consumed. The palladium and titanium foils appear to change how fast the deuterium-deuterium reaction runs without being consumed either.
The mechanism is not yet understood. The known baseline here is electron screening, a smaller effect in which electrons in the host metal partially shield the repulsion between nuclei. The new boost is several orders of magnitude larger and covers a much wider energy range, and Persaud's group is still working out what the metal lattice is doing to the deuterium. The result was reproduced across multiple foils and multiple loading conditions, and the same data appear in both the Nature Communications paper and the arXiv preprint posted in December 2025.
The application the team is pointing at is not a power plant. It is a compact neutron generator. Fusion reactions in this energy range produce neutrons that can be used to image tumors, screen cargo containers for smuggled material, and simulate the radiation environment inside a spacecraft or a planetary surface. Today, most of those neutrons come from small particle accelerators or radioactive sources that are bulky, expensive, or both. A solid foil that produces the same yield with a much smaller machine would change the engineering trade-off for those tools.
The result also lands inside an existing program. Berkeley Lab's ATAP Division leads an ARPA-E Low Energy Nuclear Reactions project, a Department of Energy effort to systematically test whether host materials can do more than survive low-energy fusion. The new paper is a milestone in that program, not an isolated curiosity. Low-energy nuclear reaction research has a long and controversial history, and the program's stated purpose is to bring data and peer review to a field where both have been thin.
The next step is mechanism. The plateau is reproducible. The explanation is not. Until the team, and the field, can name what the metal lattice is doing to the deuterium nuclei, the result is a clean, peer-reviewed data point that a long-standing assumption is wrong, and a hint that the right host material might matter as much as the right fuel.