Frozen Neon Qubit Shows Noise 10 to 10,000 Times Lower Than Typical Semiconductors

The quantum computing industry has treated noise and fabrication complexity as inseparable tradeoffs: the quieter your qubits, the more engineering they require, and the more that engineering costs. Build a billion-dollar cleanroom, run extreme ultraviolet lithography, and maybe your qubits stay coherent long enough to be useful. A paper published April 30 in Nature Electronics suggests that entire premise may have been wrong — not because the engineering got better, but because somebody tried a different material.

The work comes from Argonne National Laboratory, the University of Notre Dame, the University of Chicago, Harvard, Northeastern, and Florida State University — a coalition that pairs national lab infrastructure with academic physics programs that lack access to advanced semiconductor fabrication facilities. It is not a new discovery. The platform — electrons trapped above a frozen neon surface, controlled with microwave pulses — was first demonstrated in 2022 and set a coherence record in 2024 at roughly 0.1 milliseconds, nearly a thousand times longer than prior semiconductor qubit records. What the new Nature Electronics paper adds is the most systematic accounting yet of why the platform is so quiet: the neon surface is chemically inert and nearly impurity-free, suppressing the random electrical fluctuations, known as charge noise, that disrupt semiconductor quantum dots.

The noise characterization, performed at Argonne's Center for Nanoscale Materials, measured noise levels 10 to 10,000 times lower than conventional semiconductor qubits — competitive with the best superconducting systems, Quantum Computing Report noted. "By thoroughly characterizing the qubit's noise properties, this latest study shows why its performance is so good," said Xu Han, Argonne scientist and co-corresponding author, in an Argonne press release.

The fabrication story is where it gets interesting for anyone who builds or invests in quantum hardware. A semiconductor qubit fabrication line requires extreme ultraviolet lithography machines — each costing more than $150 million — and yields qubits that still suffer from charge noise. The electron-on-neon platform uses a resonator to trap electrons above self-assembling frozen neon; the electrons come from a filament not fundamentally different from those in incandescent light bulbs. The paper notes: "electrons are freely available from light bulb filaments, and the fabrication process is significantly simpler than traditional fabrication methods." No lithography. No cleanroom.

Neon sidesteps that trade-off — not through better engineering but through a material that is inherently cleaner. The quantum hardware industry has treated noise and fabrication complexity as a package deal: quieter systems require more engineering, which requires more money, which requires more scale. The bottleneck was never fabrication complexity. It was the wrong material solving the wrong problem.

Dafei Jin, who led the research and is now an associate professor at Notre Dame, acknowledged the platform still has limits. The team found residual noise from stray electrons and unevenness in the neon surface — both described as mitigatable through follow-up work. The platform also requires cooling below 100 millikelvin, as superconducting systems do. What the paper demonstrates is that the remaining problems are engineering problems, not physics problems.

If the electron-on-neon platform scales, the competitive advantage in quantum hardware shifts from cleanroom access to materials science expertise. The companies and national programs that built qubit strategies around silicon and superconducting fabrication face an architectural competitor that does not require their investments to matter.

The noise characterization paper does not demonstrate a working quantum computer, multi-qubit entanglement, or error-corrected operation. The 10-to-10,000-times quieter figure spans a wide range depending on frequency and measurement conditions, and independent replication has not yet been reported. But it is the most complete explanation yet of why this platform performs as well as it does — and the answer is, in part, that it is built from something almost absurdly simple.