The pursuit of a nuclear clock has all the features of a quest: a grand idea, a unique treasure, a number of false sightings, a controversy, a long period of (seemingly) fruitless searching and, finally, a pathway toward the ultimate goal.

What is a nuclear clock? The basic principle behind any type of clock is a periodic event, defining one clock 'tick'. In addition, one needs a device that can be tuned to the periodic signal, as well as a procedure for counting the clock ticks. Our present definition of a second is based on a perfect, naturally occurring oscillator: an electronic transition between particular energy levels of a caesium atom. Electronic transitions in other atoms, with much higher frequencies (4 × 1014–1.3 × 1015 Hz) than the caesium transition (9 × 109 Hz), lying in the optical range and excited by an ultra-stable laser, now form the basis of even more precise atomic clocks, which will neither lose nor gain 1 second in 15 billion years1. The nuclear clock would be based on a transition between the energy levels of a nucleus rather than electronic energy levels. A great benefit of this concept is its reduced sensitivity to the external perturbations that cause systematic errors in present atomic clocks, such as electric and magnetic fields, electric field gradients and blackbody radiation — these could be avoided or at least suppressed in a nuclear clock2,3.

Alas, nuclear transition frequencies are typically many orders of magnitude higher than those accessible by lasers! But here comes our treasure — a unique exception: a long-lived nuclear transition that occurs between an excited state (isomer) of the 229Th isotope and the corresponding nuclear ground state (pictured). Its frequency is expected to be between 150 and 170 nm, within reach of modern lasers. This transition eluded scientists during years of efforts to achieve its laser excitation; partly due to the large uncertainty in the transition frequency and its long lifetime, which is expected to be about 104 seconds. Even the very existence of the 229mTh isomer came into question after unconfirmed and disputed sightings2, but it was finally definitively confirmed4 in 2016. Experiments to verify and pinpoint the wavelength of the transition from the isomer to the nuclear ground state are ongoing2. The hyperfine structure (splitting of the atomic energy levels due to the interaction of electrons with nuclear electric and magnetic moments) of the isomer was recently observed in another breakthrough experiment5 also providing further insights into the isomer's fundamental nuclear properties.

These recent experimental milestones4,5 have provided a solid base for progress toward the realization of an actual nuclear clock. Two different types of nuclear clock design have been proposed, one based on trapped ions and another on solid-state systems2. The former involves a single trapped Th atom, with one or more electrons removed, and technologies similar to those used in current trapped-ion clocks1. This scheme offers unprecedented suppression of systematic shifts of the measured clock frequency, and an estimated clock performance with a total fractional inaccuracy of 1 × 10−19 (ref. 3).

In the other, completely different proposal, thorium is implanted in crystals transparent to radiation in the vacuum-ultraviolet range of the 229Th nuclear clock transition, such as calcium fluoride. An advantage of such a scheme is the possibility to have a large number of Th nuclei (well beyond 1010) in a solid, resulting in a significant gain in signal-to-noise ratio and clock stability2.

The development of ever more precise and stable clocks is of great importance for many purposes1, ranging from practical applications in geodesy to testing cornerstones of modern physics, such as the constancy of fundamental constants and Lorentz invariance. The variation of fundamental constants is an element of many theories beyond the standard model of particle physics and might be caused by the so-called dark matter6 — the 'missing' matter in the Universe. It has been suggested7 that the clock transition in the 229Th nucleus is enormously (by five orders of magnitude) more sensitive to the variation of the fine-structure constant and quark masses than all atomic clocks currently in operation. This result is presently a matter of controversy, with widely disparate theoretical estimates. A way to resolve this problem and to determine the sensitivity has recently been proposed, involving measurements of the nuclear quadrupole moment and the nuclear charge radius of the 229mTh isomer7. First experimental results have been just reported5.

The quest for the nuclear clock continues and the next decade may see the exploitation of a nuclear transition for new levels of metrological accuracy as well as insights into our Universe.