A Clock That Keeps Its Own Time: The First Self-Stabilizing Nuclear Clock Is Ticking in Vienna

On October 8, physicists at TU Wien in Vienna switched on a machine that does something no clock has ever done: it keeps its own time. The prototype ran for more than 24 hours with no human intervention and no conventional atomic clock supervising it. Its reference lives inside the atomic nucleus, in thorium-229 nuclei embedded in a crystal, and a feedback loop held the laser on frequency because the nuclei told it where to be.

The device, reported this week in Nature by a team led by Thorsten Schumm of TU Wien and Ekkehard Peik of PTB, Germany's national metrology institute in Braunschweig, is the world's first self-stabilizing nuclear clock. In 2024 the same collaboration showed thorium nuclei could be used as a timekeeper, but only while coupled to a conventional optical atomic clock. This prototype closes the loop: the nuclear transition itself now supplies the correction signal that keeps the laser steady. "What you really want is a self-stabilizing nuclear clock," Schumm said in the university's announcement.

The clock that corrects itself

Every clock is a counter plus a reference. In ordinary atomic clocks, the reference is a transition between energy states of an atom's electrons: a laser is locked to that transition and counted. In the Vienna device, the reference is a transition inside the nucleus of thorium-229, an isotope with an unusually low-energy nuclear state that vacuum-ultraviolet light can reach.

At the heart of the clock sits a crystal containing thorium atoms, produced at TU Wien. A laser irradiates the crystal, and the thorium nuclei absorb that light only when the laser frequency is exactly right. If the frequency drifts, even slightly, absorption decreases measurably, and the system detects the drop and readjusts the laser automatically. An optical cavity supplies short-term laser stability while the nuclear measurements correct longer-term drift; the team tested feedback intervals of 20 seconds and 30 minutes to balance correction speed against measurement noise.

Getting there required a change in detection strategy. The earlier apparatus used a pulsed laser and read the nuclear response by waiting for excited nuclei to emit fluorescence, a roughly 10-minute process that made rapid corrections impossible. Working with the Max Born Institute in Berlin, PTB built a continuously emitting ultraviolet laser source. Instead of waiting for emitted light, the team measured how much laser light the nuclei absorb, a signal that arrives immediately and makes a continuous feedback loop possible.

Twenty-three years in the making

The idea of a thorium nuclear clock was proposed in 2003 by Ekkehard Peik and Christian Tamm, but finding the exact laser frequency took the better part of two decades because the transition energy was initially known only approximately. In April 2024, the teams led by Schumm and Peik excited thorium nuclei with laser beams for the first time, demonstrating that the long-suspected nuclear transition could be driven directly.

That autumn, the team showed the thorium apparatus could indeed be used as a high-precision clock by coupling it to a conventional optical atomic clock, establishing a precise relationship between nuclear and electronic frequencies. But that device was not yet the self-stabilizing kind: it relied on the atomic clock to stay honest. The new Nature paper is the step the field had been waiting for, with a companion spectroscopy study establishing the absorption measurements and the clock study building the feedback loop on top.

The crystal approach is itself a deliberate bet. Rather than trapping isolated atoms or ions in vacuum, as the most advanced laboratory clocks do, the Vienna design embeds a large population of reference nuclei in a solid at room temperature, at the cost of material effects the team is still learning to control.

Precise, but not yet record-breaking

Measured over the course of a day, the prototype reached a fractional precision of approximately 10 to the power of minus 15, equivalent to an error of roughly one second in 30 million years. Within a continuous run, the fractional frequency instability sat in the low 10 to the minus 14 range, verified by comparison against an ytterbium-ion optical clock at Austria's Federal Office of Metrology and Surveying over a stabilized fiber link.

The headline is not that a nuclear clock beats atomic clocks, not yet. The headline is that a transition inside the atomic nucleus, ten thousand times smaller than the atom around it, has held a laser steady for a full day, unattended.

The honest caveats matter as much as the numbers. Between runs, frequency reproducibility was limited to approximately 5 times 10 to the minus 13: realigning the laser before each run could send the beam through different regions of the crystal, where the resonance center varied by up to 1.7 kilohertz, possibly from local strain caused by structural irregularities, though that remains a hypothesis. Measurements at the same crystal position reproduced cleanly, so the clock is stable within a run and shifts between them.

The team's stated roadmap is concrete: stronger ultraviolet lasers to cut measurement noise, longer light paths through the crystal, more uniform crystals and a reproducible beam path to fix the between-run drift, and possibly alternative host materials with narrower resonances. The best optical atomic clocks remain more precise, but for a first prototype of an entirely new class of timekeeper, the Vienna result is a foundation, not a finish line.

Two teams, one week

The same issue of Nature carries a second thorium-229 nuclear clock from a research group in Beijing, the second working device in the world after Vienna and Braunschweig. Both teams used thorium-doped calcium fluoride crystals as a solid host, arriving at the same architecture independently.

That parallel announcement matters more than the race framing suggests. Independent construction of the same device within the same week is how a field moves from theory to engineering: the result is reproducible, the components are understood, and more than one laboratory can now push the technology forward. In about a week, nuclear timekeeping went from a single-lab demonstration to a small but real international field.

Beyond timekeeping

Why build a clock around the nucleus at all? Because the nucleus is more than 10,000 times smaller than the atom, it is far less disturbed by electromagnetic fields and other environmental influences than the electron cloud, making nuclear clocks potentially more robust and more sensitive probes of fundamental physics.

The Vienna team has already put that sensitivity to work. Comparing a nuclear clock against electronic clocks tests whether fundamental constants drift over time, and can search for hypothetical ultralight dark matter particles that would cause periodic shifts in the transition frequency. The researchers analyzed roughly 23 hours of data taken with 20-second feedback intervals, found no oscillation above their detection threshold, and placed upper limits constraining proposed dark-matter models, with the caveat that those constraints depend on the assumed nuclear sensitivity, which is not yet precisely known.

This is not a device that will appear on anyone's wrist. It is metrology infrastructure: the kind of instrument that underpins the definition of the second, navigation and communications networks, and tests of the constancy of nature's constants. The thorium crystal in Vienna is ticking, and for the first time, nothing else is keeping it honest.