Inside the incredibly slow race to reinvent time


ANDREW LUDLOW’S is no ordinary ticker. An intricate tangle of tubes, cables and lasers occupying an entire room at his lab in Boulder, Colorado, it is one of the best timekeeping devices ever made. “It’s the Lamborghini of atomic clocks,” he says.

That isn’t to say it is fast. But Yb-2, as the clock is known, is precision engineered. In fact, it should measure out each passing second so precisely that it wouldn’t miss a beat for around 20 billion years – more than the age of the universe.

This is the stunning frontier of precision at which timekeeping now finds itself. Clocks such as Ludlow’s could spur on as yet unheard-of technological innovations. They could transform our understanding of the universe, revealing wrinkles in established laws of physics and variations in the fundamental constants of nature that would otherwise be impossible to detect. But for metrologists like Ludlow, they raise an even more fundamental question: is it time once again to redefine time?

That might like seem an odd thing to consider for what is a fundamental property of the universe. The flow of time is an enigma; many physicists even suggest it is just an illusion. But clock time is our own invention. We define its basic units – the hours, minutes and seconds that break up the day. They started out as subdivisions of the time it takes Earth to rotate around its axis. Indeed, when astronomer Christiaan Huygens invented the pendulum clock in the 17th century, a second became firmly established as 1/86,400 of a solar day, a factor derived from the division of the day into 24 hours, then 60 minutes per hour and finally 60 seconds per minute.

But Earth isn’t a dependable metronome. The duration of its rotation varies by microseconds daily and progressively slows ever so slightly, meaning a second gradually gets longer. That became a problem in the early 20th century, when experimental verification of quantum mechanics and the emergence of radio broadcasting required a steadier, more precise unit of time. It eventually arrived with the microwave atomic clock: a timepiece that ticks in harmony with the frequency of microwave radiation emitted from the rapid oscillations inside caesium atoms, where electrons hop back and forth between closely spaced energy levels.

The first microwave atomic clock was unveiled at the National Physical Laboratory (NPL) in Teddington, UK, in 1955. It was accurate to 1 second every 300 years, meaning two such clocks would fall out of sync by just one second every three centuries. It wasn’t long before such precision transformed the way we measure the basic unit of time. In 1967, representatives at the 13th General Conference on Weights and Measures in Paris officially redefined the second as “the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom”.

The new second was no longer or shorter than the old one. But the change did provide a much more precise definition of that duration, dramatically improving the extent to which we can be sure each second is the same as the next, and the one after that too.

It is a time standard that persists today, even if the accuracy of microwave atomic clocks has improved to the point that the best caesium clocks now keep time with an accuracy of 1 second in roughly 300 million years. And it is a measure that has served us well. The steady backbeat of the caesium atom’s vibration underpins all manner of modern technologies, from GPS and smartphones to the internet and electricity grids, all of which require exquisitely precise synchronisation.

But it is no longer the best we can do – not by a long way. Ludlow, a physicist at the National Institute of Standards and Technology’s (NIST) Physical Measurement Lab, is one of many researchers working on optical atomic clocks, a new generation of timepieces that promise to once again dramatically improve the precision with which we can measure time passing.

We have known for a long time that other atoms oscillate much faster than caesium. Strontium and ytterbium stand out because the electrons surrounding their nuclei have stable excited states, relatively unperturbed by potentially disruptive outside forces like temperature and electric and magnetic fields. The problem was always that their electrons transition between energy levels so fast that there was no easy way to count them.

That wrinkle was smoothed out in 1999 with a device called an optical frequency comb, which essentially translates the atomic oscillations measured in the optical range into microwave frequencies. For the first time, the rates at which optical clocks “tick” could be calibrated against one another and the standard set by the caesium atom.

“The current definition of a second is no longer the best we can do”

The technique sparked something of an arms race, with labs around the world competing to create ever more precise optical clocks. Currently, top contenders include not only Ludlow’s ytterbium clock at NIST, but also a similar device at the RIKEN Quantum Metrology Lab in Tokyo and strontium clocks at NIST and the National Metrology Institute of Germany in Braunschweig.

These optical clocks have already achieved a level of certainty nearly two orders of magnitude higher than caesium-based clocks, to the point that most would lose a second only once over the course of the entire history of the universe.

That might seem like overkill. Wrist watches and iPhones don’t need to operate to a precision of 18 digits or more. Yet if they can be made sufficiently portable, optical atomic clocks could be used for all sorts of practical purposes, from tracking movement to detecting volcanic activity and earthquakes. They are also likely to usher in many technologies and breakthroughs that we haven’t thought of yet – and that is before we even get onto the fundamental questions in physics that more accurate timepieces would help resolve (see “Clocking new physics”).

“It’s an ‘If you make it, they will come’ kind of attitude,” says Anne Curtis, a metrologist at NPL. “Fifty years ago, when people were contemplating GPS satellites for the first time, no one thought we’d be walking around with handheld computers utilising GPS in real time, just to get to a restaurant.”

So what are we waiting for? If optical clocks have already achieved such record-breaking precision, why isn’t the world already ticking to their superior beat? “Frankly, it is somewhat awkward to have optical clocks that are ‘better’ than the very definition of what a second is,” says Franklyn Quinlan at NIST. The trouble is that there is a checklist of knotty problems to be addressed before they can establish a new international time standard.

For one thing, you need to pair the signals coming from optical atomic clocks with all the electronic infrastructure already in place, which is currently synchronised using microwave-based clocks. That is tricky because, as well as an optical frequency comb, it requires a separate piece of hardware called an optical-to-electrical converter to transform pulses of light into an electric signal. For a long time, it wasn’t clear that it was possible to translate the exquisite timing produced by optical clocks into the microwave frequency range for use in electronics.

But earlier this year, Quinlan, Ludlow and their colleagues cracked the problem. After a decade of work, they finally demonstrated that the translation provided by optical frequency combs yields microwave signals with a 100-fold stability improvement compared with the best microwave atomic clocks. “Considering that it took 20 years of steady improvements to see the last tenfold increase in microwave signal stability, we think a sudden 100-fold increase is a significant advance,” says Quinlan.

Curtis agrees. “As part of the roadmap for the redefinition of the second, it is an essential requirement to be able to connect the future optical definition back to the current microwave definition,” she says. “This demonstrates this at the highest levels.”

Metrologists are now discussing the idea of submitting a proposal to vote on an official redefinition at the next General Conference on Weights and Measures, scheduled for 2026.

To make that happen, the field must first negotiate a few obstacles. For starters, metrologists will have to decide on the cut-off point, a level of precision that everyone agrees is enough for redefinition. “People start getting antsy about how much better do you need to be than the [current] definition before you should redefine it,” says Curtis.

Final countdown?

Once that’s settled, they will have to thrash out which kind of optical atomic clock should be used to set the official redefinition. There are at least 10 different models being developed in labs around the world, and no single candidate has yet emerged as an obvious best choice.

The clocks differ not only in the types of atoms being used, but also in their architectures. One leading design, the optical lattice clock, measures the oscillations of about 10,000 neutral atoms simultaneously to provide stable, snappy readings. NIST’s Yb-2 clock is an example of this. Other candidates include the single-trapped-ion clock, which measures the transition frequency of a single isolated, charged atom in a way that can help to reduce uncertainty. The problem with single-atom clocks is that they deliver a smaller signal than lattice clocks and thus take more time to produce measurements.

Fritz Riehle, former head of optics at the National Metrology Institute of Germany, says that the diversity in clock designs is a good thing for now, because it provides different possible solutions. But eventually, one must be crowned the winner. This decision will ultimately fall to the board of representatives at the General Conference on Weights and Measures; its conclusion will be based on the recommendation reached by numerous experts, committees, working groups and subgroups. Naming a final winner will probably be more of a human problem than a scientific one, says Riehle, although one he is sure “will be solved in a competitive but respectful way”.

“Nature places a fundamental limit on how well we can measure time”

Before we get to that point, though, there are still scientific hurdles to clear – not least the verification of the various measurements that the clocks produce. This process ensures consistency and replicability, and it is used to compare optical clocks with each other and the best microwave clocks. Teams in Colorado, France, Germany, the UK, Italy and Japan have already begun using optical fibres to link optical clocks to facilitate such comparisons. But labs still sometimes produce slightly different results, leaving researchers troubleshooting.

According to Curtis, this is all par for the course. “The art and science in metrology is really about assessing all the things that can go wrong,” she says. “There’s no reason why we can’t, over the next five years, all figure out what might be going wrong with our clocks and ensure that it doesn’t.”

How long a new definition of the second will last is anyone’s guess. Just like their microwave-based predecessors, optical atomic clocks will, at some point, be surpassed. In fact, people are already thinking about clocks based on transitions that take place inside the nucleus of an atom, rather than the cloud of electrons that orbit it. “We want to choose something that will last for a long period,” says Patrizia Tavella, director of the time department at the International Bureau of Weights and Measures in France. “But we also understand we can’t have something that will last forever.”

Even future generations of atomic clocks will eventually have to grapple with the nature of time as described by Einstein’s general theory of relativity, which predicts that a clock ticks ever so slightly faster for every centimetre it is elevated within Earth’s gravitational field. As the precision of our best timepieces reaches into ever more digits, shifts in that field will begin to interfere. At some point, we will run up against “the fundamental limit nature places on us in how well we can measure time”, says Jun Ye, another NIST physicist.

It is most likely, then, that a timeless definition of the basic unit of time, just like time itself, will always escape us., 19 August 2020
; ”>