Time is what clocks do

Since 1967, the second has been defined by an atom. Specifically, it is the time taken for 9,192,631,770 cycles of the microwave radiation that flips a cesium-133 atom between two energy states. Cesium clocks built on that definition are the backbone of global timekeeping, satellite navigation and the internet’s synchronization. They are also, by the standards of the best laboratories, starting to look a little slow.

Faster pendulums

Every clock counts the swings of something that oscillates regularly. A grandfather clock counts a pendulum; a quartz watch counts the vibrations of a crystal; a cesium clock counts microwaves. The more swings per second, the finer you can divide time.

Optical clocks use atoms such as strontium, ytterbium or aluminum, probed with visible or ultraviolet laser light instead of microwaves. Strontium’s transition oscillates about 429 trillion times per second, nearly 47,000 times faster than cesium’s. The best of these clocks have systematic uncertainties of around one part in 10¹⁸, about a hundred times better than the cesium standards. Over the 13.8 billion years since the Big Bang, an error of that size accumulates to less than half a second.

Counting that many cycles was itself a problem, since no electronics can follow a signal at hundreds of terahertz. The solution, the optical frequency comb, translates optical frequencies down to ones that electronics can count, and it earned a share of the 2005 Nobel Prize in Physics. Optical clocks had to wait for it.

Raise a clock by a centimeter and it runs faster by about one part in 10¹⁸.

Gravity in a millimeter

Einstein’s general relativity says clocks run slower closer to a massive body. Near Earth’s surface, the fractional change in rate is given by a simple expression: the acceleration of gravity times the height difference, divided by the speed of light squared. Raise a clock by a centimeter and it runs faster by about one part in 10¹⁸.

It is worth watching that experiment shrink over time. In 1959, Robert Pound and Glen Rebka detected gravity’s effect on the frequency of gamma rays sent down a 22.5-meter tower at Harvard. In 2010, researchers at the National Institute of Standards and Technology saw the effect between two optical clocks separated in height by about a third of a meter. In 2022, a team at JILA in Colorado measured the difference in ticking rate between the top and bottom of a single cloud of strontium atoms about a millimeter tall. At that scale the expected shift is about one part in 10¹⁹, and the measurement agreed with relativity.

One consequence is that there is no longer any such thing as the time at a laboratory. There is the time at a particular height. Someone who spends an eighty-year life on the top floor of a 100-meter building ages about 28 microseconds more than a twin at street level. Harmless, but entirely real, and now measurable in a benchtop instrument.

Rewriting the second, and retiring the leap second

Two different changes to timekeeping are under way, and they are easy to confuse. The first concerns how long a second is. Metrologists plan to redefine it in terms of optical transitions, with an international roadmap holding to 2030 as the decisive date. The open questions are which atom to choose, or whether to define the second using a weighted ensemble of several, and how to show that optical clocks around the world agree well enough to share a standard.

Comparing them is its own challenge. Satellite links are not precise enough, so laboratories have connected clocks with optical fiber across hundreds of kilometers and carried portable optical clocks between countries. Gravity complicates even this, since a clock’s height above sea level must be known to within centimeters before its rate can be compared with another’s.

The second change concerns how the calendar stays in step with Earth. On 15 October 2026, the General Conference on Weights and Measures voted to stop adding leap seconds to Coordinated Universal Time from 20 May 2027. Earth’s spin is irregular and has lately been speeding up a little, raising the prospect of a first negative leap second that few computer systems were built to handle. Under the new rules, atomic time and the sky may drift apart by as much as an hour before anyone has to act. The clocks, in effect, have stopped checking with the planet.

What precision buys

Better clocks are not only about timekeeping. Because gravity changes a clock’s rate, a network of optical clocks could measure height differences of a centimeter across a continent, a technique called relativistic geodesy that could track sea-level rise and the movement of magma beneath volcanoes. Clocks are also sensitive probes of new physics: if fundamental constants drift, or if certain kinds of dark matter pass through Earth, the ratios between different atomic clocks might change. Precision has a habit of forcing such questions, as the long saga of the muon’s magnetism shows.

Further out still is a clock based on the nucleus of an atom rather than its electrons. In 2024, researchers finally excited a long-sought low-energy transition in the thorium-229 nucleus with a laser, and linked it to a strontium clock. A nuclear clock would be less sensitive to its surroundings and potentially more sensitive to changes in the laws of physics themselves.

Einstein is often quoted as saying that time is what clocks measure. The physicists building optical clocks might put it more strongly. Time, as we can know it, is what clocks do, and the clocks are getting very good.

Sources

• Bothwell et al., “Resolving the gravitational redshift across a millimetre-scale atomic sample”, Nature, 2022
• BIPM: Frequently asked questions on the redefinition of the second
• Euronews: Leap second to be scrapped from 2027
• Tiedau et al., “Laser Excitation of the Th-229 Nucleus”, Physical Review Letters, 2024
• Zhang et al., “Frequency ratio of the ^229mTh nuclear isomeric transition and the ⁸⁷Sr atomic clock”, Nature, 2024