Conventional atomic clocks are essentially optical counters: they rely on spectroscopy — a method for measuring the frequency of radiation with high precision as electrons jump between energy levels. But thorium-229 has a special feature: its nucleus has an excited state with a half-life of 641 seconds. In nuclear physics, that's a pendulum in thick syrup — one swing lasts an eternity, yet its rhythm is impeccably stable. It's this nuclear transition, not an electronic one, that underpins the next generation of clocks.

When these nuclei are embedded in a crystal lattice, a tricky problem emerges. Lattice defects, microscopic inhomogeneities, act like an uneven floor: each nuclear 'pendulum' swings in its own local field, throwing off its rhythm. The physicists' task is to level that 'floor' by suppressing the influence of defects. If they succeed, the clock's accuracy would reach a fantastic limit: an error of one second over the entire age of the Universe. Such clocks could detect gravitational time dilation, the difference in the flow of time between floors of a building, and also stress-test the Standard Model of particle physics — the theoretical framework describing fundamental interactions in the search for new physics.