The Hard Part of a One-Nanometer Superconductor Was Keeping It Alive

Atomically thin superconductors have had a strange problem. Their physics can be extraordinary, but simply taking them out of the chamber where they are made can begin destroying the material. Niobium diselenide, or NbSe2, can remain superconducting when thinned to a single layer, yet that layer oxidizes rapidly in air. For quantum devices, the obstacle has therefore been less about discovering superconductivity than making the material survive normal fabrication. A team led by researchers at MIT, NYU and other institutions reported a different manufacturing approach in Nature in August 2026. Instead of growing monolayer NbSe2 first and trying to cover it before oxygen causes damage, they place graphene on a silicon-dioxide substrate before the superconductor is formed. Chemical precursors then move through a gap less than one nanometer thick and grow the NbSe2 underneath the graphene. The graphene performs two jobs at once. It helps guide the growth into a smooth, continuous monolayer, and it immediately becomes a protective cap. The researchers produced films more than an inch across, transferred them without exposing the NbSe2 to oxidation, and connected the material to a conventional superconducting microwave circuit. The resulting monolayer remained superconducting at roughly 1 kelvin and showed about 0.7 nanohenries per square of kinetic inductance after integration. That last property is important for quantum hardware. Kinetic inductance comes from the inertia of the superconducting charge carriers themselves. A material with high kinetic inductance can provide a large inductive response in a very small area, which is useful for compact resonators, detectors and circuit elements. A separate 2026 Nature Communications study measured the effect across few-layer NbSe2 and found it increased as the material became thinner, reaching about 1.2 nanohenries per square in the monolayer limit. This is not a high-temperature-superconductor breakthrough. One kelvin is still extremely cold, and today's quantum computers already rely on mature superconducting materials and fabrication processes. The new work also does not prove that monolayer NbSe2 will outperform those materials in complete qubits. Contacts, microwave loss, fabrication yield and long-term reliability will matter as much as the impressive material properties. The more important achievement is manufacturability. Two-dimensional materials are often demonstrated as tiny flakes with unusual behavior, then become difficult to scale because the same material is fragile, contaminated or inconsistent outside a laboratory experiment. Encapsulation epitaxy changes that sequence by building the protection into the growth process itself. That makes this less a story about finding a new superconductor than about turning a delicate quantum material into something engineers can actually process. If the method extends to other air-sensitive two-dimensional materials, graphene may matter not because it is the active device, but because it lets the material underneath survive long enough to become one.