Get ready for the new, improved second

Modern civilization, it is said, would be impossible without measurement. And measurement would be pointless if we were not all using the same units.

>>Alanna MitchellThe New York Times
Published : 2 May 2022, 06:27 AM
Updated : 2 May 2022, 06:27 AM

So, for nearly 150 years, the world’s metrologists have agreed on strict definitions for units of measurement through the International Bureau of Weights and Measures, known by its French acronym, BIPM, and based outside Paris.

Nowadays the bureau regulates the seven base units that govern time, length, mass, electrical current, temperature, the intensity of light and the amount of a substance. Together, these units are the language of science, technology and commerce.

Scientists are constantly refining these standards. In 2018, they approved new definitions for the kilogram (mass), ampere (current), kelvin (temperature) and mole (amount of substance). Now, with the exception of the mole, all of the standards are subservient to one: time.

The metre, for example, is defined as the distance light travels in a vacuum during one-299,792,458th of a second. Likewise, the new definition of the kilogram rests on the second, in a manner too complicated to explain in fewer than several paragraphs.

“All the units now are not autonomous units, but they are all depending on the second,” said Noël C. Dimarcq, a physicist and the president of the BIPM’s consultative committee for time and frequency.

That means that conceptually, if clumsily, you could express other units, such as weight or length, in seconds.

“You go to the grocery and say, ‘I would like not 1 kilogram of potatoes, but an amount of seconds of potatoes,’” Dimarcq said.

Yet now, for the first time in more than a half-century, scientists are in the throes of changing the definition of the second, because a new generation of clocks is capable of measuring it more precisely.

In June, metrologists with the BIPM will have a final list of criteria that must be met to set the new definition. Dimarcq said he expected that most would be fulfilled by 2026, and that formal approval would happen by 2030.

It must be done carefully. The architecture of global measurement depends on the second, so when the unit’s definition changes, its duration must not.

“It’s like a once-in-every-50-year thing,” said Elizabeth A. Donley, chief of the time and frequency division of the National Institute of Standards and Technology, or NIST, in Boulder, Colorado. She is on the BIPM’s international consultative committee with Dimarcq. “And so it’s a big deal that we want to get right, and so there’s a lot of discussion. It’s exciting to work on, for sure.”


Once, humans told time by looking at the heavens. But since 1967, metrologists have defined time instead by measuring what’s going on inside an atom — clocking, as it were, the eternal heartbeat of the universe.

But time still has its roots and even its nomenclature in astronomical time keeping. Originally, it was based on the path of Earth in its daily spin, day to night and back again. Eventually, ancient Egyptian astronomers who used the duodecimal counting system, based on 12, divided the day and night into 12 hours each, giving us 24 hours in the day.

Those hours varied in length, depending on where Earth was in its orbit around the sun. A little more than 2,000 years ago, Greek astronomers, who needed fixed hours to calculate things like the movements of the moon, developed the revolutionary idea that a single day ought to be divided into 24 hours of the same length.

That same astronomical thinking led them to patch the ancient Babylonian method of counting by 60, the sexagesimal system, onto the hour. Just as they divided a circle or the sphere of Earth into 60 parts, and then 60 again — making 360 degrees — so they divided the hour.

The first division of the day’s 24 hours (known in Latin as partes minutiae primae) gave them the length of the minute, which was one-1,440th of an average solar day. The second division (partes minutiae secundae) provided them the duration — and name — of the second, which was one-86,400th of a day. That definition stood, in effect, until 1967. (There was a brief detour into something called ephemeris time that was so complicated even metrologists did not use it.)

But the definition had problems. Earth is gradually slowing in its daily rotation; days are growing slightly longer and so the astronomical second is, too. Those small differences add up. Based on extrapolations from historical eclipses and other observations, Earth as a clock has lost more than three hours over the past 2,000 years.

Therefore, the standard unit of time, based on astronomical reckoning, is not constant, a reality that became increasingly intolerable for metrologists during the first decades of the 20th century as they discovered just how irregular Earth’s spin was. Science demands constancy, reliability and replicability. So does time — and by the late 1960s, society was becoming increasingly reliant on the frequencies of radio signals, which demanded extremely precise timings.

Metrologists turned to the far more predictable movement of atomic particles. Atoms never wear out or slow down. Their properties do not change over time. They are the perfect timepieces.

By the middle of the 20th century, scientists had coaxed atoms of cesium 133 into divulging their secret inner ticks. Cesium, a silvery-gold metal that is liquid at about room temperature, has heavy, slow atoms, which means they are relatively easy to track.

Scientists put cesium atoms in a vacuum and exposed them to the energy of microwaves, in the nonvisible range of the electromagnetic field. The task was to figure out which wavelength, or frequency, would excite as many cesium atoms as possible into emitting a packet of light, or photon. The photons were picked up by a detector and counted.

The wavelength that won the contest was designated as the natural frequency resonance of the atom. Think of it as a pendulum operating in a rhythm unique to that type of atom.

In the case of cesium 133, the frequency is nearly 9.2 billion ticks per second — 9,192,631,770, to be precise. The length of the second used in the experiment was based on the length of the day in 1957 when the original scientific experiments were taking place, and was derived from measurements of Earth, the moon and stars. By 1967, metrologists at the BIPM had set the natural frequency resonance of cesium 133 as the official length of the second.

Despite that cesium-based definition, astronomical time and atomic time are still inextricably conjoined. For one thing, atomic time occasionally needs to be adjusted to match astronomical time because Earth continues to change its pace at an irregular rate, whereas atomic time remains constant. When atomic time gets nearly one second faster than astronomical time, the timekeepers stop it for a moment, allowing Earth to catch up — they insert a leap second in the year. So while the duration of the second does not change, the duration of a minute occasionally does. After an initial insertion of 10 leap seconds in 1972, timekeepers now add a leap second to atomic time roughly every year and a half.

In addition, as weird as it may seem, we still tick through 1957-era seconds, even with our modern atomic clocks. That is because the natural frequency resonance of cesium 133 was measured in 1957 and locked to the duration of the astronomical second in that year, a fact that will not change even when the second is redefined once more.


The redefinition is in the works because scientists have developed new instruments called optical atomic clocks. These operate on similar principles to cesium clocks but measure atoms that have a much faster natural frequency resonance, or tick. Those frequencies are in the visible, or optical, range of the electromagnetic spectrum, rather than the microwave range, hence the name.

There are several species of optical clock, each counting the ticks of a different atom or ion — ytterbium, strontium, mercury, aluminium and more. So far, no species has emerged as the clear favourite for the upcoming redefinition.

“Optical clocks are very definitely not ready for prime time,” said Judah Levine, a physicist at NIST’s time and frequency division. “They are laboratory projects.”

For one thing, although they are built to examine such tiny atoms, most are massive, about the size of a heavy dinner table. Some fill a laboratory. They are also difficult to operate.

“It requires a whole bunch of specialists who are chained to the table, if you know what I mean,” Levine said. “It’s not just push a button and walk away.”

In all, about 20 or 30 optical atomic clocks of all species exist today, Donley said.

Three are in Boulder. A typical one is settled on a steel slab to isolate it from floor vibrations. It is shielded from disturbances in Earth’s magnetic field. At its heart is a vacuum chamber about a foot in diameter, containing whichever atom or ion is under scrutiny. Some clocks contain a single ion. Others contain thousands of the same type of atom.

Lasers are mounted on the sides of the table. They chill the atoms or ions to near absolute zero, trapping them in place and slowing them down. Then the lasers probe the atoms or ions, beaming a nearly pure colour of light on them that scientists tune to find the precise wavelength that will elicit the desired tiny shift in energy.

“Just as a child only achieves great height on a playground swing if her parent’s pushes arrive at the right rhythm, the atoms become detectably excited only if the laser colour is tuned perfectly,” Jeffrey A. Sherman, a physicist in NIST’s time realisation and distribution group, explained in an email.

The trick is then to be able to read the laser’s colour to determine the precise frequency of the wave that elicits the shift in energy. And this is where the optical atomic clock’s secret weapon kicks in. A key component of the clock is a second type of laser called a femtosecond-laser frequency comb, the discovery of which led to a Nobel Prize in physics in 2005. It is a pulsed laser, equivalent to a series of spikes of light spaced by precisely the same amount, like the teeth of a hair comb.

This comb of light can read the wavelengths of the pure-colour lasers that are exciting the atoms or ions. The waves are fast, moving at rhythms, or frequencies, some 100,000 times that of the microwave energy that excites cesium. This enables optical atomic clocks to measure time far more precisely than cesium clocks.


Why do we need such precision? Partly because time is not just time; it is tied to, and influenced by, gravity and mass. Nor is time constant, despite what the existence of an international standard might suggest. Albert Einstein’s theory of general relativity, for example, predicts that time moves more slowly when it is near a massive body, such as a planet, because it is slowed by gravity’s pull.

That means that if the tick of a clock changes, even very slightly, the physical conditions in which the clock is situated may have changed, too. Being able to read these changes opens the possibility that the clocks could detect such entities as dark matter or gravitational waves, Donley said.

“They’re very exquisite tests of fundamental physics, which is one of the exciting things about optical clocks,” she said.

One experiment has already taken place. In 2015, physicists at NIST were in the early days of developing their optical atomic clocks. They were puzzled by the fact that the seconds were measuring slightly differently across the clocks, which were in labs spread throughout Boulder.

Then they thought about the theory of general relativity. Could these optical clocks be responding to slight changes in gravity?

They asked Derek van Westrum, a physicist at the National Geodetic Survey, which is part of the National Oceanic and Atmospheric Administration, to investigate. In 2015 and 2018, van Westrum measured height differences among the labs where the clocks were stationed. Like time, height is linked to gravity and mass.

His traditional survey levelling techniques, which measure height above sea level, found that the clocks were indeed at different heights. Their slightly different measurements of time were capturing minuscule changes in the gravitational field. A clock just 1 centimetre higher than another ran faster.

“That Einstein’s crazy prediction of what mass and gravity does to time would actually have a practical application, to me is just incredible,” van Westrum said, chuckling.

If several optical atomic clocks could be placed in different parts of the world, geodesists could measure ticking differences between them, and therefore differences in height and the gravitational field, he said. For example, a network set up near a flooding river could explain where the water would flow and identify escape routes for residents.

Such possibilities lie in the future. Today, physicists are still trying to make optical clocks talk to one another over distances, an imperative for time keeping. Optical clocks cannot communicate efficiently over satellite systems, for example, because satellite time keeping is not yet optical.

Physicists are making strides. A recent experiment at NIST, published in Nature last year, linked the three clocks in Boulder through both optic fibre and air.

And scientists are looking once more to the heavens for help. Now, though, it is not to track the movements of planets or stars, but to use information from far beyond our galaxy.

Using very long baseline interferometry, researchers in Italy and Japan recently tried to link two optical atomic clocks about 5,500 miles apart. The experiment involved several antennas reading radio signals from distant outer space, and then linking the information to atomic clocks.

It worked, and for a moment time and space merged, mediated by the stars.

©2022 The New York Times Company