A thorium nuclear clock ran itself for 24 hours, drifting a second every 30 million years

A thorium-229 nuclear clock built at TU Wien in Vienna has a fractional precision of about one part in 10 to the 15th power, the equivalent of roughly one second of error over 30 million years. The result matters less for the number than for the design: the clock corrects itself, which the earlier thorium experiments, still tied to a conventional atomic clock, could not do.

Timekeeping of this kind underpins GPS, phones and the internet, and nuclear clocks could eventually outdo atomic clocks.

A laser, a crystal and a feedback loop

According to a report on the university’s announcement, the clock kept itself on frequency for more than 24 hours with nobody adjusting it. Thorsten Schumm, a professor at TU Wien’s Institute of Atomic and Subatomic Physics who leads the team, said in the university’s news release that the goal was a nuclear clock that stabilizes itself, and he described the basic idea as one laser plus thorium. The thorium sits inside a special crystal grown at TU Wien, and the laser is aimed at it. He added that, in principle, “much higher precision is possible” when nuclei rather than atoms do the timekeeping, because nuclei are more than 10,000 times smaller than atoms and so are less affected by outside disturbances.

The nuclei of thorium-229 have two energy states separated by an unusually small gap, so a laser can push them between the two. They absorb light only at one exact frequency. If the laser drifts away from that frequency, absorption falls, and the system detects the drop and nudges the laser back. That loop is what makes the clock self-stabilizing, and the release says it can work without relying on a conventional atomic clock as a reference.

The work builds on two earlier steps. In April 2024 the Schumm group and the team of Ekkehard Peik at the national metrology institute PTB in Braunschweig showed that thorium nuclei could be excited with a laser, confirming a long-suspected nuclear transition. In autumn 2024 the thorium setup was coupled to a conventional optical atomic clock, which kept the laser steady while the nuclei served as the timekeeper. The release stresses that this was not yet a self-stabilizing nuclear clock, whereas the new design is described as able to run without one.

What 24 hours and 30 million years actually measure

The two headline figures come from the same experiment, described in the Nature paper titled “A thorium-229 optical nuclear clock with feedback loop,” whose lead author is L. Toscani De Col. According to the release, the clock ran steadily for more than 24 hours without intervention, and the precision of about 10 to the minus 15 was measured over a single day.

That second point deserves care, because the shorthand is easy to misread. The comparison to one second in 30 million years is an extrapolation of the one-day stability, not the outcome of a test lasting millions of years. A separate summary of the two papers, from martincid.com, puts the Vienna clock’s fractional instability at about 3 times 10 to the minus 12 over one second, shrinking toward 10 to the minus 15 as the averaging time reaches a day. The runtime figure is the only longer-term claim made for the device. The team is open about where the prototype stands. Its precision, the release says, “is not yet at the level of the world’s best optical atomic clocks.” Schumm’s own verdict was that “for a first prototype it is a fantastic result.”

A second clock in Beijing and what comes next

Vienna was not alone. A team at Tsinghua University led by Shiqian Ding reported a nuclear clock of its own in the same Nature issue, titled “A nuclear clock synchronized to 229Th” according to the martincid.com summary. Ding’s group built its own continuous laser at 148.4 nanometres, more powerful than the Vienna system, and grew small calcium fluoride crystals doped with thorium-229 from just 1.4 micrograms of the isotope, since access to thorium-229 in China is limited.

Two independently grown crystals gave the same transition frequency, which also matched earlier measurements at JILA in Colorado. Citing Physics World, the same summary says the Beijing clock’s one-second instability is about six times lower than Vienna’s, though it gives no absolute figure for Beijing. Ding said the two teams worked independently and that “the concept is robust.”

Both groups name the crystal as a limit. Ding said calcium fluoride host crystals restrict performance and that better host materials are needed, and he called a clock built on single trapped thorium ions “technically much more difficult.” For Vienna, the release lists stronger lasers and higher-quality thorium crystals as the planned improvements, framed as future steps rather than results.

What none of the sources settle is how long such a clock can run beyond the 24-hour test, or what it would cost and weigh outside a laboratory. The release gives no figures on either. Whether the one-day stability holds up over weeks, and whether the Beijing clock’s lower instability translates into a better long-run number, will have to come from the next round of measurements by Schumm’s group, PTB’s Peik and Ding’s team in Beijing.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview