Two teams, working thousands of miles apart, have just shown that a long-sought new kind of clock can run. One, built in Beijing, kept a steadier beat in early tests. The other, in Vienna, ran for a full day and used its device to look for dark matter. The papers appeared in Nature on the same day. What the numbers actually show, and what they do not, is the part that matters.
A Clock That Uses the Atom’s Core
Most precise clocks today watch electrons. These tiny particles orbit the nucleus of an atom. When light hits them at just the right color, they jump to a higher energy level and drop back. The speed of those jumps becomes the clock’s tick.
A nuclear clock tries something harder. It watches the nucleus itself—the dense center made of protons and neutrons. Nuclei are far smaller than the electron cloud around them. They also sit deeper inside the atom, so outside electric and magnetic fields disturb them less. That is the hope: a steadier reference that is harder to knock off course.
Only one known nucleus makes this practical right now. Thorium-229 has two energy states so close together that a laser can drive the jump between them. The needed light sits in the vacuum ultraviolet range, near 148 nanometers. Scientists first proved they could excite this transition with a laser in 2024. That opened the race. By early October 2026 both the Chinese and European groups had closed the loop: they locked a continuous laser to the nuclear transition and treated the result as a working clock.
What the Chinese Team Actually Built
The Beijing group is led by Associate Professor Ding Shiqian at Tsinghua University, working with the Beijing Academy of Quantum Information Sciences and partners at several Chinese institutes. They published “A nuclear clock synchronized to 229Th” in Nature on 7 October 2026.
They grew their own calcium fluoride crystals doped with thorium-229. Because the isotope is scarce, they used only 1.4 micrograms—roughly 200 times less than the amount the European team employed. From that tiny starting amount they produced millimeter-scale crystals. They also built their own continuous-wave vacuum-ultraviolet laser at 148.4 nanometers, generating about 10 microwatts of power. Earlier work by the same group on that laser had already appeared in Nature in February 2026.
Light from the laser passes through the crystal. A detection system watches how the nuclei absorb the light. When the laser drifts off the target frequency, a feedback loop corrects it. In the published tests the clock reached a fractional frequency instability of 5 × 10⁻¹³ divided by the square root of the averaging time in seconds. Two crystals grown separately agreed with each other at the 10⁻¹³ level. Those readings also matched earlier measurements made with different crystals and different equipment at JILA in Colorado.
Tsinghua’s own summary notes that the nucleus is less open to outside disturbance than the electron shell. The team says the result turns a decades-old idea into a working frequency reference and gives a solid-state platform that can be reproduced.
The Vienna Device and the Direct Comparison
The European effort is led by Thorsten Schumm at TU Wien in Austria and Ekkehard Peik at the German national metrology institute PTB. Their paper, “A thorium-229 optical nuclear clock with feedback loop,” also appeared in Nature on 7 October 2026.
They used higher concentrations of thorium in their calcium fluoride crystals and ran the device continuously for more than 24 hours without hands-on correction. During that run they compared it with a ytterbium-ion atomic clock. The short-term instability figure reported for their system is about 3 × 10⁻¹² per square root of averaging time. That is roughly six times larger than the Chinese figure, which is why secondary reports describe the Tsinghua clock as having a sixfold stability gain in the early tests.
Over a full day the Vienna clock approached a fractional instability near 10⁻¹⁵. One way to picture that is an error of roughly one second in 30 million years. The team also used the clock to search for ultralight dark matter by looking for tiny periodic shifts or slow drifts in the nuclear transition energy. They set limits but did not detect a signal. Those limits sit at a level comparable to the best existing atomic clocks.
Schumm has described the situation as a “fierce but friendly” competition. He has also pointed out that the two approaches have complementary strengths: the Vienna crystals contain more thorium and have good optical properties, while the Beijing laser is more powerful. Combining the advantages, he said, could produce a significantly better clock.
What the Numbers Do and Do Not Mean
Short-term stability and long-term accuracy are not the same thing. The Chinese device shows lower scatter over short averaging times in the published data. The Vienna device demonstrated continuous unattended operation for a day and a direct comparison against a mature atomic clock. Neither has yet reached the performance of the best optical atomic clocks now running in laboratories. Those can accumulate an error of only about one second over tens of billions of years.
Reproducibility is a separate practical question. If every crystal shifts the nuclear frequency in a slightly different way, each clock would need its own calibration. The Chinese team’s finding that two separately grown crystals agreed at the 10⁻¹³ level addresses that worry. It suggests the host crystal does not unpredictably scramble the signal.
Both groups stress that the technology is still experimental. Improvements in laser power, crystal quality, and isolation from the environment are expected to raise performance by large factors. Peik has noted that the natural linewidth of the transition should be far narrower than what has been observed so far; reaching that limit would open the door to still better stability.
Why Anyone Outside a Physics Lab Might Care
Better clocks already matter for navigation satellites, financial time stamps, and power grids. Nuclear clocks are not ready to replace those systems. If they can be made compact and robust, however, they could one day move high precision out of specialized laboratories. Ding has pointed to possible uses in navigation, space-based measurements, and precise synchronization of networks.
They also offer a new tool for testing the basic rules of physics. Because the nuclear transition is unusually sensitive to certain constants, a nuclear clock compared against an ordinary atomic clock can look for slow changes in the strength of the electromagnetic force or the strong nuclear force. Papers from 2007 and 2008 already mapped how such changes would appear in the thorium-229 transition. The Vienna dark-matter search is an early example of that kind of use.
Time itself is not only a laboratory number. The length of a day on Earth changes by tiny amounts because of shifts deep inside the planet. A recent study on gravitational torque from the solid inner core, discussed on this site, shows how even small measured effects matter for satellites and leap seconds. See Earth’s Inner Core Quietly Changes the Length of a Day. Tools that can resolve still smaller fractions of a second give researchers another way to separate the planet’s motion from the clocks that measure it.
The Wider Picture of the Work
China’s path was shaped by limited access to thorium-229. The team responded by developing a laser and a crystal-growth method that use very little material. That independent route produced a working clock at the same moment as the European effort. Both results appeared in the same issue of Nature, which is why the milestone is being described as shared even while the short-term stability numbers differ.
Chinese laboratories have been moving quickly on other advanced hardware as well. A plant in Liuzhou that aims to finish one industrial humanoid robot every ten minutes is one recent example covered here: China Humanoid Factory Goes Live: Robots Building Robots in Liuzhou. Precision measurement and quantum devices sit in the same broad push.
Neither group claims a finished product. Both describe next steps: stronger lasers, better crystals, longer continuous runs, and tighter control of temperature and vacuum conditions. Schumm has said the nuclear clock is still far from its target performance. Ding’s team has said it will keep refining laser stability and the vacuum systems around the crystal.
Independent researchers have called for further checks of the published numbers. That is ordinary in a new field. The fact that two separate teams reached working devices with different hardware, and that one of them showed crystal-to-crystal agreement, is what makes the result hard to dismiss as a single-lab curiosity.
What Remains Open
The best optical atomic clocks still set the standard for absolute accuracy. Nuclear clocks start with a theoretical advantage in robustness and in sensitivity to certain physics effects, but they have not yet turned that advantage into a measured record. Supply of thorium-229 remains tight; the isotope is also of interest for medical research, so commercial quantities will not appear overnight.
Whether the sixfold short-term stability edge holds after longer runs and independent verification is the next clear question. Whether either design can be shrunk into something that fits on a satellite or in a field instrument is a later one. The papers supply the numbers and the methods. The rest is engineering and time.
The race that began when the first laser excitation of the thorium nucleus was shown in 2024 now has two operating clocks and a public record of how they compare. That is the concrete advance. The larger claims about navigation, fundamental constants, and the future of timekeeping will be settled by the devices that follow these prototypes.
Original source: Reporting draws from the Nature papers published 7 October 2026 (Huang et al., A nuclear clock synchronized to 229Th; Toscani De Col et al., A thorium-229 optical nuclear clock with feedback loop), the Tsinghua University physics department summary of 9 October 2026, and contemporaneous accounts in Reuters, the South China Morning Post, and Physics World. Secondary summaries such as the one carried by NaturalNews on 11 October 2026 rest on those primary reports.
Disclaimer for fact-checkers: The performance figures and dates above are taken from the peer-reviewed Nature abstracts and the institutions’ own statements. Short-term stability comparisons are those reported in the literature; long-term accuracy claims remain projections. No commercial product is available.