Seven clocks, four countries, one fiber cable
A fiber cable runs from a laboratory in Paris to the National Physical Laboratory in Teddington, 785 kilometres west, threading through countryside and under the seabed of the English Channel. In February 2026, researchers at LNE-OP in Paris sent an atomic clock signal through it. At the other end, the NPL team had a clock of their own running independently — measuring different atoms, built by different scientists. For 65 days, the two signals ran in parallel, and the two labs recorded how well they matched.
They agreed better than anyone had verified before, across any international border.
The campaign was coordinated by Marco Pizzocaro of INRIM in Turin and involved 43 colleagues at four national metrology institutes: LNE-OP in Paris, NPL in Teddington, PTB in Braunschweig, and INRIM in Italy. Seven optical clocks participated. LNE-OP contributed a mercury-199 optical lattice clock — one of only a handful operating on that transition in the world, using ultraviolet light at 363 nanometres that most labs find too difficult to work with. NPL and PTB each ran ytterbium-ion traps. INRIM and NPL added strontium and ytterbium lattice clocks. The fiber links connecting them ran 1,023 kilometres from Paris across the Alps to Turin, 1,370 kilometres northeast to Braunschweig, and 785 kilometres west under the Channel to London.
The sharpest result came from the two ytterbium-ion clocks at NPL and PTB. Built independently, never previously directly compared, they agreed in their measured frequency ratio to within 7.7 × 10⁻¹⁸ (Physical Review Research). The practical translation: two clocks run from the beginning of the universe to the present — 13.8 billion years — and arrive at today within roughly one second of each other. No two independently developed optical clocks had ever confirmed each other at this level of precision across national borders.
The difficulty is not purely technical. Several labs can build a single clock accurate to 10⁻¹⁹. The harder problem is agreement: different scientists, different atoms, different continents, one fiber cable — and you need the answer to be the same. Not all comparisons in the campaign were clean; PTB’s strontium lattice clock had to be excluded from part of the analysis due to a frequency offset from its clock laser, and some frequency-ratio measurements showed scatter beyond expected noise. The fiber link itself contributed negligibly to the final uncertainties. The disagreements, where they existed, were in the physics (INRIM).
What the campaign demonstrated — and what the International Committee for Weights and Measures requires before redefining the second — is that the same frequency ratio can be recovered independently, by different groups, to 18 digits, across the continent. The current SI second, defined in 1967 by the hyperfine transition of the caesium-133 atom, went through the same gauntlet: a single measurement by Louis Essen at NPL in 1955, confirmed and cross-checked for twelve years at multiple labs, before it was adopted as the global standard. Optical clocks are now doing the same work, over the same institution, through the same building in Teddington.
The target for a new definition is around 2030. The European campaign did not resolve everything. But it showed that the ytterbium-ion clocks agree. The fiber works. The network is real.
Sources
- International Optical Clock Comparison Using the European Optical Fiber Network — arXiv:2604.27963 — primary source; campaign timeline, participating institutions, clock species, fiber distances, Yb⁺ agreement at 7.7×10⁻¹⁸, exclusions from analysis
- Europe links the clocks that could redefine the second — INRIM — INRIM press release on the significance of the multi-lab comparison and the path toward SI redefinition
- Louis Essen — Wikipedia — context on the 1955 NPL caesium measurement and the 1967 SI definition of the second