The Race to Build the World’s Most Precise Clock

Barrett’s clock, located at the university’s Centre for Quantum Technologies, is intended to slice time into more and smaller segments than any clock before it. An analog stopwatch can typically divide the second into 10 pieces: 0.1 second, 0.2 second, etc. The lutetium clock will, in theory, add 14 zeros to the right of that decimal point, thereby segmenting the second into (roughly) a quadrillion pieces, and remain accurate to within a second if left running continuously for 30 billion years. Even the smallest vibrations from, say, the exhaust fans that keep the laboratory environment clean are enough to upset that precision—and with it, all the lab’s work.

So far, the potential applications for a clock with 15 decimal places of precision are theoretical and aspirational. But that’s not stopping labs around the world from competing, using different elements, to reach the same goal. Lutetium is in some ways the least likely element to feature in a winning effort. Not only is it relatively rare in nature; because of its high cost and similarity to other elements, its known applications are also rare. Thus researchers and industry tend to pass it over. If Barrett and his lab succeed, the global market for lutetium won’t expand significantly—after all, the clock uses hardly any. Nevertheless, lutetium will quickly become the beating heart of science and the global economy.

All clocks measure time by following a recurrent event. That can mean tracking the rotation of the Earth via the shadows on a sundial, or it can mean counting the swings of a pendulum. None of these timekeeping methods can produce a measurement that’s constant in all times and places, though. Temperature and humidity alter mechanical clock parts, lengthening or shortening seconds. Timekeeping based on astronomical observation is hampered by the irregularity of the Earth’s orbit around the sun.

The bulk of human history so far hasn’t required much precision, so the limitations of clocks were never really a problem. Then in the 19th century, atomic physics opened up the possibility of a universal, unchanging definition of time. The idea was deceptively simple. Each atom of an element on the periodic table vibrates uniquely when bombarded by an exact frequency of radiation. So if the number of cycles of radiation necessary to stimulate an atom can be counted during the best existing non-atomic definition of a second, the world would have a new definition of the second grounded in the unchanging atom. In effect, it’d have an atomic clock.

Over a period of almost two decades, scientists determined that cesium, whose vibrations are comparatively easy to observe and which exists in only one stable form, and thus wouldn’t need to be purified, was the best element to use. In 1967, the General Conference on Weights and Measures, an intergovernmental organization, replaced the old celestial definition of a second with the amount of time it takes to measure 9,192,631,770 oscillations of the microwave required to stimulate vibration of cesium-133.

Even before the atomic second was defined, there were scientific, military, and commercial ventures deeply interested in taking advantage of more precise timing. The U.S. military was one of the earliest patrons of atomic clocks, and in 1973 the armed forces introduced the global positioning system, a super-accurate navigation network that uses satellites outfitted with atomic clocks to measure how long it takes a signal to reach and bounce back from a receiver on Earth.

Missiles and other military technologies were the immediate beneficiaries of GPS. The private sector benefited, too, from Google Maps to modern telecommunication networks that synchronize using GPS. According to a June 2019 study sponsored by the National Institute of Standards and Technology (NIST) in Colorado, GPS was responsible for $1.35 trillion in economic benefits to the U.S. from 1984 to 2017. China, Europe, and Japan have their own satellite navigation systems relying on atomic clocks; a collective outage, as unlikely as that is, would cripple the global economy.

For nearly two decades scientists around the world have been working on improved atomic clocks that stimulate atoms using visible light, which oscillates roughly 100,000 times faster than the microwaves in a cesium clock. Creating these optical clocks isn’t just a matter of switching out microwaves for lasers and cesium for another element. Instead, scientists must overcome a range of scientific and technical challenges, including the fundamental question of which element is best suited for the work of telling time.

There’s no lack of candidates, including aluminum, mercury, and strontium. For now, the front-runner appears to be ytterbium, which is being investigated by large, well-funded teams at NIST and labs in Europe and Asia. In late 2018, NIST announced that a pair of ytterbium-based clocks had set records for precision, stability, and reproducibility of results. Lutetium was a late entry to this intense and expensive competition. Yet, in a relatively short time, it’s proven itself a worthy and intriguing dark horse, thanks to Murray Barrett.

Barrett didn’t aspire to be a clockmaker. After completing his Ph.D. at Georgia Tech in 2002, where he studied atomic physics, he did a two-year postdoctoral fellowship at NIST in the quantum computing and information program. He returned briefly to New Zealand before taking up his current position in Singapore. His research initially focused on what he calls “quantum information stuff,” a field that marries computing to the tiny world of quantum physics. According to Barrett, he was in his office one day and “just happened to look at lutetium and got an interest in precision measurement.” He pauses at the skepticism on my face, then adds: “It really is as arbitrary as that.”

Even temperature changes so minute they wouldn’t register on earlier generations of atomic clocks can be a source of error for super-sensitive optical clocks. Lutetium, it turns out, is uniquely insensitive to temperature shifts. “And I would argue,” Barrett says, “whatever atom has the least sensitivity to its environment will make the better clock.”

Standing in his lab, surrounded by home-built lasers and electronics, some of which sketch ghostly red lines in the darkness, Barrett invokes two rulers of similar length as a way of explaining why a more accurate clock is necessary. The rulers, he says, can be thought of as clocks. The cesium ruler measures in centimeters; but an optical clock based on lutetium “would measure in millimeters,” and therefore allow the world to be measured with greater accuracy. It might even allow the measurement of things that could never be measured before, thereby revealing and resolving fundamental questions of physics.

For example, Albert Einstein’s general theory of relativity posited that the faster you travel, the more gravity pulls on you and the more time slows down. To verify this, scientists have flown atomic clocks in space and found that they ran slower, relative to clocks on Earth. Similarly, an atomic clock on a mountaintop runs faster than one at the mountain’s base, where gravity is stronger. If optical clocks could be shrunk to portable size, something Barrett and others are researching, those measurable shifts in time could be used to map the Earth’s topography in unprecedented detail. In theory, they might also be able to reveal mineral deposits beneath that topography.

The scientific mother lode from optical clocks would be measurements so precise they could reveal dark matter, the mysterious substance that makes up about 27% of the universe’s mass. Its existence has only been inferred, because ordinary matter alone can’t explain the evolution and structure of massive objects such as galaxies. NIST’s ytterbium-based clocks are reaching levels of precision that may eventually prove sufficient to detect signatures of dark matter.

In Boulder, Andrew Ludlow, the project leader on the record-setting NIST clocks, isn’t celebrating yet. “Right now it’d be difficult to imagine a full-fledged optical clock developed at a price point it could be deployed all over,” he says. Making matters even more difficult, the intricacy of optical clocks means they’re subject to significant downtime. For them to serve as references for a new second, much less detect dark matter signatures, they’ll need to be able to run uninterrupted, 24/7. “Because of the complexity of these clocks—and the cost—a lot of the applications are situations where you don’t need a lot of them, but you need them to be really good under special conditions,” Ludlow says. “Various military applications fall into that category.” Ludlow’s funders include NASA and the Defense Advanced Research Projects Agency, better known by its acronym, Darpa.

But a market for optical clocks is percolating. “I’ve been contacted by a couple of people in the financial industry with generic curiosity about optical clocks,” Ludlow says. An optical clock could potentially allow an exchange to time-stamp and order many more trades in ever smaller increments of time, a quality that would be of particular interest to high-frequency traders and those who like to foil them. In 2016 the hedge fund Renaissance Technologies filed a patent on an exchange-trading system that relies upon “atomic clocks, optical clocks, quantum clocks, Global Position System (GPS) clocks, or any clock capable of measuring time, accurate within microseconds.”

Barrett smiles uncomfortably when I ask him about the financial industry’s interest in optical clocks. He’s seated at his desk in a spacious office across from his lab, a whiteboard covered in arrows and equations beside him. He didn’t get into physics to find ways to time-stamp more trades. Instead, the search for precision “suits my personality,” he says, and he’s devoted to improving it through the endless tinkering necessary to make any clock better. “The hard part is knowing when to stop and publish,” he says.

Barrett happily acknowledges that for now his lab and lutetium aren’t the leaders in the race to build the second-defining optical clock. For one thing, he isn’t nearly as well-funded as other groups. But what he lacks in resources is made up for by the properties of “my atom,” he says. “My argument is, if you can make an accurate clock in ytterbium, then you must be able to make an accurate clock in lutetium, and it should be more accurate because the properties are just better.”

Singapore’s Agency for Science, Technology and Research, a government office that sponsors R&D that could benefit the city-state competitively, is helping bankroll him. “Why do they fund it?” Barrett asks. “If you look at the history of clocks, as we make them better and better, new applications pop or the ones we have get better and better.”

To anyone outside the clock labs, the first evidence of progress will be a new definition of a second. The soonest that can happen, officially, is the next gathering of the General Conference on Weights and Measures, scheduled for 2022. But most observers believe there won’t be a global consensus on which element or clock design should serve as the basis for a new second until at least the 2026 meeting. For now, Barrett, like other clockmakers, remains focused on the ticking. “That’s what you really do in the clock business,” he says. “You ask, ‘How do we make these measurements better and better?’ ”
 
This story is from Bloomberg Businessweek’s special issue The Elements.

©2019 Bloomberg L.P.