Two clocks that keep time using an energy jump inside the thorium-229 nucleus are now running in two independent labs. A European team led by Luke Toscani De Col of the Vienna Center for Quantum Science and Technology published its results on 3 June 2026. A Chinese team led by Beichen Huang of Tsinghua University followed on 7 June. Both studies are preprints and have not yet been peer reviewed. The Vienna prototype ran for more than 24 hours without human help, but it is still less stable than the best optical atomic clocks.

The Austrian members of the European team come from the Vienna University of Technology, and they worked with the German institute PTB in Braunschweig.

A jump in the nucleus instead of vibrating electrons

Today's atomic clocks count time by the vibration of electrons around the nucleus. Nuclear clocks use a jump inside the nucleus itself, between two of its energy states. The nucleus is more than 10,000 times smaller than the atom and is far better shielded from outside disturbance, such as electromagnetic fields and changes in temperature.

Thorium-229 is a rare exception. Its jump needs so little energy that an ultraviolet laser can trigger it, while other nuclei need far more powerful sources. Both teams embed thorium-229 nuclei in a calcium fluoride crystal and shine this light on them.

The clock works through feedback. Professor Thorsten Schumm of the Vienna University of Technology says the laser changes the state of the nuclei, and the nuclei in turn keep the laser's frequency steady. The nuclei absorb light only at one exact frequency. When absorption drops, the laser has drifted, and the system pulls it back. So the clock does not need an ordinary atomic clock to give it a reference. The Vienna team measures absorption continuously and gets an instant response, rather than waiting for a faint trace of light after the nuclei are excited.

Stability of about one part in a quadrillion

The Vienna clock has a stability of about one part in a quadrillion (10-15) when measured over a day. That equals an error of about one second in 30 million years. Before that, the system's short-term noise was about three parts in a trillion. In the Chinese team's work, the instability of the laser locked to the nucleus comes close to one part in 10 trillion after one day. The two teams also differ in approach: the Chinese use a more powerful laser, while the Austrians use a crystal with a higher thorium concentration.

Schumm admits: "This is not yet the level of the best optical atomic clocks, but for a first prototype it is a fantastic result." A stronger laser and a better crystal will bring improvements, and the plan is to use them in the next prototypes.

The biggest problem is the crystal. When the laser is aimed again between two series of measurements, it hits a different spot in the crystal and the results differ. As a result, day-to-day repeatability is about five parts in 10 trillion. The authors blame local strain in the crystal. If it can be brought under control, improved solid-state versions could reach about one part in 10 quadrillion and come close to the best optical atomic clocks, while staying small and fairly simple.

Dark matter search with 23 hours of data

The Vienna team used about 23 hours of data to look for dark matter in the form of ultralight scalar fields. Such a field would change the clock's frequency. The team found no signal, but its limits on a link to the strong nuclear force are 100 to 1,000 times tighter than in earlier experiments with atomic clocks.

The same principle can be used to test whether the constants that govern the forces of nature are truly constant. Researchers also expect uses in navigation, including for distant spacecraft, and in mapping gravity underground for geophysics.