- Singapore’s CQT reports two lutetium-ion clocks with record-evaluated uncertainties of 1.2 and 1.3 × 10⁻¹⁹, making them the most accurate atomic clocks to date
- The stated results top Wuhan’s calcium clock (4.4 × 10⁻¹⁹) and NIST’s aluminum clock (5.5 × 10⁻¹⁹) by a wide berth
- It lands days before the CGPM meets in Versailles, where a draft resolution admits no consensus on replacing cesium just yet
A team at the National University of Singapore (NUS) has put a single charged atom of lutetium at the top of the rankings for the world’s most accurate clock.
The remarkable achievement was documented in a paper published in Nature, as researchers at NUS’s Centre for Quantum Technologies (CQT) reported two lutetium-ion clocks with evaluated fractional uncertainties of 1.2 × 10⁻¹⁹ and 1.3 × 10⁻¹⁹, the lowest yet reported for any atomic clock.
“I am confident that what we have now is the most accurate clock in the world,” team leader Murray Barrett said in the NUS announcement.
A significant leap in accuracy and why it matters
Computing the numbers, 1.2 × 10⁻¹⁹ corresponds to one second of error in roughly 264 billion years, consistent with reports that it would take more than 260 billion years to lose a second in accuracy.
That is about 3.7 times lower than the previous record, a liquid-nitrogen-cooled calcium-ion clock from the Chinese Academy of Sciences in Wuhan that reported 4.4 × 10⁻¹⁹ in February 2026, and 4.6 times lower than the aluminum-ion clock NIST unveiled in July 2025 at 5.5 × 10⁻¹⁹.
The feat is doubly impressive given the size (and therefore relative resources) of the team that recorded the breakthrough, underscoring how close a clock can come to a “true” time figure in terms of accuracy.
The CQT group built two independent clocks and compared them over 200 hours using correlation spectroscopy, which cancels noise from their shared laser. They agreed to within 5.7 × 10⁻¹⁹, which the team calls the most precise clock comparison ever made.
“A man with two watches is never sure,” joint first author Kyle Arnold said in CQT’s release, quoting an old saying to argue that the only real test of a standard is comparing clocks and showing they reproduce each other.
The group has worked on lutetium for more than a decade and believes it is the only team building clocks with it. It is also significantly more practical than the next most accurate alternative: Wuhan’s calcium clock needed its ions’ surroundings chilled to about 80 kelvin with liquid nitrogen to suppress blackbody radiation.
Singapore’s clocks ran at room temperature in what the paper calls “relatively basic linear Paul traps without magnetic shielding,” using commercially available lasers. Barrett says the clock would keep its performance anywhere from Death Valley’s record heat to the frozen Antarctic plateau.
The real prize is the definition of the second, which has rested on cesium since 1967. The best optical clocks now beat cesium’s realization by up to 100 times, and the old yardstick has become the bottleneck that many are looking to replace with their own solution.
The General Conference on Weights and Measures, which approves changes to the SI, meets in Versailles from October 13 to 15 2026, and has published Draft Resolution B, which picks no successor. It concedes “a consensus has not been reached” on whether to use one atomic species or several, or which, flags “noticeable inconsistencies” in some optical clock comparisons, and asks for a proposal at the 2030 meeting.
Singapore’s emphasis on verification speaks to that gap, but lutetium is a newcomer on a long list of contenders. The team’s next step, according to joint first author Michael Lee, is “to take the lab-scale clock and miniaturize it into a transportable system.”
That matters, because a clock that can travel might be the practical way to check it against other labs at such a microscopic level. One could therefore argue that the most important test of Singapore’s record has not happened yet.
