Atomic Clocks
The Most Stable Tickers Known to Humanity
Why Atoms Tick Better Than Anything Else
Every clock works by counting something that repeats. A pendulum clock counts swings. A quartz watch counts vibrations of a tiny crystal. A wall clock at home is good to a fraction of a second per day. A quartz watch is good to a fraction of a second per week. But every mechanical or electronic oscillator drifts. The pendulum wears at its pivot. The crystal slowly changes shape with temperature and age. There is always something physical inside the clock that ages, and aging means drift.
Atoms have no such weakness. Every cesium atom in universe is identical to every other cesium atom. Every transition between two of its internal energy states happens at exactly the same frequency, today, a thousand years ago, and a thousand years from now. There is nothing inside an atom that can age or drift or shift with temperature, because there is no internal mechanism in the human-engineering sense. The atom's tick is set by the quantum mechanics of how its electrons sit around its nucleus, which depends only on the values of the fundamental constants of nature. As long as those constants stay constant, the atom keeps perfect time.
This is the simple idea behind atomic clocks. Build a machine that locks an electronic oscillator to a chosen atomic transition, and you get a clock whose long-term stability is set by quantum mechanics rather than by any engineering tolerance. The best such clocks in 2026 lose less than one second over the entire age of universe so far. They are the most precise measurement devices humanity has ever built, and they have quietly become indispensable to physics, geodesy, navigation, and basic infrastructure.
How an Atom Becomes a Clock
An atom can sit in any of a discrete set of internal energy states. Each state has a definite, sharp energy. If an atom in its lowest state absorbs a photon of exactly the right frequency - one whose energy matches the gap to a higher state - the atom jumps to the higher state. If you tune the photon's frequency away even slightly, nothing happens. The atom is exquisitely picky.
A clock turns this pickiness into a stable reference. The machine generates radiation at an adjustable frequency, shines it on a sample of atoms, and watches how many of the atoms absorb it. If the frequency is exactly right, many absorb. If it drifts away from the transition frequency, fewer absorb. A feedback loop continuously adjusts the radiation source to keep the absorption maximum. The result is that the radiation source is locked to the atomic transition frequency. Count cycles of that radiation source, and you have a clock whose pace is anchored to physics rather than to any mechanical part.
The Cesium Definition of the Second
The first generation of practical atomic clocks ran on cesium. The cesium-133 atom has a particular transition between two nearly identical low-energy states - a "hyperfine" splitting caused by the interaction of the outer electron's tiny magnetic moment with the nucleus's tiny magnetic moment. The transition frequency is about 9.2 gigahertz, which falls in the microwave band. Microwaves are easy to generate, easy to detect, and easy to count. Cesium clocks have been mature engineering since the 1960s.
In 1967, the international body responsible for measurement standards redefined the second. Before then, the second was defined astronomically, as a fraction of the day. The new definition was simply: one second equals exactly 9,192,631,770 oscillations of the cesium-133 hyperfine transition. The number was chosen to match the previous astronomical second as closely as possible. Since then, every clock on Earth, including the timestamp on this page, ultimately traces back to a chain of comparisons leading to a network of cesium atomic clocks.
The modern version is the cesium fountain clock. A cloud of cesium atoms is laser-cooled to near absolute zero, tossed gently upward through a microwave cavity, slowed by gravity, and falls back through the cavity again. The two passes through the cavity at slightly different moments form a measurement that is much more precise than a single pass. The best cesium fountain clocks are good to about one part in a thousand trillion over a day - a fractional uncertainty smaller than a single second over thirty million years. That is far better than anything humanity can do mechanically, but it is no longer the best technology available.
Optical Clocks Tick Tens of Thousands of Times Faster
A clock's precision improves with the frequency it ticks at. A higher-frequency oscillator gives you more cycles per second to count, and a smaller fractional uncertainty in any single cycle translates into a more precise overall measurement. Cesium ticks at about 9.2 gigahertz. The transitions used in modern "optical" atomic clocks tick at hundreds of terahertz - in the visible-light range, tens of thousands of times faster.
Optical clocks use atoms such as strontium, ytterbium, or aluminum. The transitions chosen have to be extremely narrow - the atom has to be picky in frequency to a level far beyond cesium - and they have to be addressable with a laser at a known optical frequency. The hard engineering problems are different from cesium fountains. The atoms have to be held nearly motionless to avoid Doppler shifts, the laser has to be exceptionally stable on its own, and the experimentalists need a way to count optical cycles, which is far above what any ordinary electronics can do.
All of these problems have been solved. The atoms are held in an "optical lattice," a regular grid of microscopic traps formed by overlapping laser beams. Each trap holds a single atom and freezes its motion. Optical-frequency counting is done using a device called a frequency comb, invented in the late 1990s and recognized with the 2005 Nobel Prize in Physics. A frequency comb effectively divides an optical frequency down to an electronically countable rate without losing precision, bridging the optical and the radio worlds.
By the mid-2020s, the best strontium and ytterbium optical lattice clocks at laboratories such as JILA in Boulder, NIST in Maryland, and PTB in Germany routinely reach a precision of about one part in a billion billion. That is a hundred times better than the best cesium clocks. Two of these clocks placed side by side will agree on the passage of one second within a billionth of a billionth.
Clocks That Disagree Over a Centimeter
General relativity predicts that clocks tick more slowly in stronger gravitational fields. A clock on the floor of a lab ticks slightly slower than a clock on a shelf in the same lab. The effect is tiny - about one part in ten million billion per meter of height near Earth's surface - and for almost all of physics history it was a theoretical curiosity that could never be directly demonstrated at human scales.
Modern optical clocks have crossed that threshold. In 2022, the JILA group demonstrated a single strontium lattice clock so stable that, by looking at slightly different vertical positions within the same atom cloud, the researchers could measure the gravitational redshift between layers separated by less than a millimeter. The clock literally ticked at slightly different rates at different heights inside a sample about the size of a fingernail. Einstein's prediction from 1916 about the effect of gravity on time is no longer something you have to fly a clock on an airplane to detect. It happens at your kitchen counter.
This precision opens an entirely new use for clocks: measuring gravity. By comparing the rate of one optical clock to another, you measure the height difference, the local gravitational potential, and the shape of Earth's gravity field in a region. This is called "chronometric geodesy," and it is becoming a practical technology. By the late 2020s, networks of optical clocks linked over fiber-optic cables across continents are expected to map Earth's gravity field at centimeter accuracy, useful for monitoring sea-level rise, ice-sheet thinning, and underground water reservoirs.
An Atomic Clock in Space
In April 2025, the European Space Agency launched the Atomic Clock Ensemble in Space, or ACES, to the International Space Station. ACES contains two complementary atomic clocks - a cesium-based clock optimized for the microgravity environment and a stable hydrogen maser - linked by laser to ground clocks on Earth.
ACES has several jobs. It tests the predictions of general relativity for clocks moving rapidly in a different gravitational potential than the ground - the same effect that GPS satellites have to correct for, now measured with new precision. It compares ground-based atomic clocks at laboratories around the world that cannot easily be connected directly by fiber, using the station as a flying clock to compare. And it provides an independent check on whether the fundamental constants of nature are drifting over the years.
Through 2026, ACES has been steadily producing data confirming relativistic predictions to better precision than ever achieved from space, and it has begun supporting the international comparison campaigns that will eventually redefine the second.
Hunting Dark Matter With Clocks
A surprising application of high-precision atomic clocks is the search for dark matter. Most direct-detection experiments look for dark-matter particles colliding with ordinary atoms, and those efforts have found nothing despite decades of trying. A different category of dark-matter candidates would not collide at all. Instead, they would be ultralight fields oscillating very slowly through space. If such a field exists and couples even weakly to the electron or proton mass, then the energy levels of every atom in universe would oscillate ever so slightly along with it. The transition frequencies of atomic clocks would oscillate too.
Comparing two atomic clocks built on different elements is the cleanest test. The relative tick rate of, say, strontium versus aluminum depends on a specific combination of fundamental constants. If a dark-matter field is jiggling those constants, the ratio will oscillate. Over the past few years, several groups have used networks of optical clocks worldwide to look for exactly this effect. So far, no detection. But each null result tightens the bounds on what kinds of ultralight dark matter can exist. By 2026, atomic-clock searches have ruled out enormous swaths of parameter space that no other technique could touch. The clocks have not found dark matter yet, but they have eliminated entire classes of candidates that other methods would have missed.
The Nuclear Clock Frontier
Every atomic clock so far uses transitions between electron energy states. The electrons sit far enough outside the nucleus that nearby electric and magnetic perturbations can shift their energy levels and limit precision. The atomic nucleus itself, by contrast, is much smaller, much more tightly bound, and harder to perturb. A clock built on a nuclear transition would in principle be even more stable than the best optical atomic clocks.
The trouble is that almost all nuclear transitions sit at energies thousands of times higher than visible light - in the X-ray range or beyond. No tabletop laser can drive them. The one known exception is thorium-229, which has an extraordinarily low-energy excited nuclear state that can be reached with a vacuum-ultraviolet laser. For decades, thorium-229 was a theoretical curiosity that no one knew how to actually excite.
In 2024, two milestones landed in quick succession. A team at TU Wien and Germany's PTB achieved the first direct laser excitation of the thorium-229 nuclear isomer, using a thorium-doped crystal. Months later, a JILA group measured that transition's frequency with a frequency comb at the precision needed to use it as a clock reference. The breakthrough opens the door to a nuclear clock - an entirely new generation of precision timekeeping. Through 2025 and into 2026, several groups have begun building nuclear-clock prototypes. They are not yet competitive with optical lattice clocks for sheer stability, but they are sensitive to a different combination of fundamental constants, making them an independent probe for new physics. A nuclear clock and an electronic clock together provide a more discriminating test of whether the constants are drifting than either alone.
A Coming Redefinition
Optical lattice clocks have been better than cesium clocks for more than a decade. The international community has nonetheless held off on redefining the second, mostly because the underlying technology was still moving rapidly and the international comparison campaigns needed to validate any new definition were not yet mature. By the late 2020s the situation is changing. Multiple optical-clock species have been compared against each other to extraordinary agreement, optical fiber and satellite links allow continent-scale clock comparisons, and ACES is bridging across regions that fiber cannot reach.
The committee that defines the second has set out conditions that must be met before the redefinition can happen. Through 2026 those conditions are being progressively met. The plan is to redefine the second in terms of one or more optical-clock transitions, most likely around the end of the decade. When that happens, the precision of every other measurement that depends on time - and most of physics ultimately does - will improve in step.
The Bigger Picture
Atomic clocks are usually mentioned as a technology - the engine inside GPS, the reference behind every financial timestamp, the standard behind every legal definition of a second. They are all of those things. But they are also one of the most sensitive measurement instruments fundamental physics has ever invented. They detect gravitational time dilation across millimeters of height. They map Earth's gravity from orbit. They search for dark-matter fields slower than universe is old. They limit how much the fundamental constants of nature can have drifted since the Big Bang.
None of this requires anything beyond what was already true a century ago: an atom has discrete energy levels, set by quantum mechanics, that do not change with time. Take that observation seriously enough, engineer around its consequences for a hundred years, and you end up with the most precise instruments humanity has ever built and a new way to probe fundamental questions about gravity, dark matter, and the constants of nature. The atom does not change. We just keep finding new uses for its refusal to drift.
