Original URL: http://www.theregister.co.uk/2011/11/21/npl_measuring_the_second/

Atomic time boffins build better second-watcher

NPL and precision on the final frontier

By Gavin Clarke

Posted in Science, 21st November 2011 13:00 GMT

US researchers recently named the atomic clock at the UK's National Physical Laboratory (NPL) in London the most accurate atomic timepiece on the planet.

Experts from the University of Pennsylvania found that NPL-CSF2 loses just one nanosecond every two months.

It's a badge of honour for NPL, the birthplace of the first successful atomic clock, whose daddy was controversial physicist Louis Essen who built Caesium I – in 1955.

Essen worked with NPL's John Parry, as senior scientist on the clock. He went on in later years to outrage the scientific establishment by claiming that errors in Albert Einstein's theory of relativity meant the formula E=mc² was wrong. Others have since followed Essen.

When it comes to clocks, the NPL-CSF2 descendant of Essen's original work is a big hitter.

Dr Krzysztof Szymaniec, leader of NPL's CSF2 project, told us in an interview: "Whoever you get your time from, they get their time from a clock that uses international time, and we are at the top of the pyramid."

NPL CSF2 is one of a six atomic clocks used to ensure the accuracy of the international standard for time, Universal Co-ordinated Time (UTC), upon which all clocks and devices rely for time.

TAI is set by 300 atomic clocks but these also slip and it is NPL CSF2 – with its five atomic peers – that keep the rest honest, providing the primacy frequency standard.

NPL CSF2 is the best of the best.

NPL CSF2 and Krzysztof Szymaniec, photo Gavin Clarke

Szymaniec and CSF2: the future is more accurate

The US study, lead by professor Kurt Gibble and published here, is providing more than an occasion for some patriotic pendulum waving, though: it's paving the way for the construction of a timepiece that will measure the second even more accurately.

NPL has started work on CSF3 – an early version is scheduled for early 2012 – and Gibble's work is helping NPL refine one component that will be important in helping overcome the effects of a force currently impeding absolute accuracy in CSF2.

That piece? Something called a cavity, which is shaped like a Coke-can and though which lasers are shot and millions of caesium atoms are streamed – the process by which physicists measure the second. It reduces distributed cavity phase, which is like Doppler effect and makes accurate readings of a second very hard to take.

You probably won't notice the impact of the cavity or CSF3 on your daily routine of getting a train on time or picking the kids up from school at the end of the day, but NPL is not looking to the present. The team hopes to lay the time-keeping foundations for a generation of future applications yet un-realised. This could include deep space flight where accurate time is critical to an understanding your exact location and where getting it can mean the difference between a soft landing in Florida or making a wrong turn and getting lost light years away. If you think that's a all a bit far-fetched, consider where we are today.

"The question of application at this stage of development is legitimate," Szymaniec concedes. "Yeah, you could say it's just for he sake of it but there will be more applications once the technology is available.

"GPS and sat nav is so popular now but only because you can put these atomic clocks on satellites and maintain the accurate time scales. Forty years ago when these [atomic] clocks were coming out, if someone said: 'Who needs an atomic clock we are happy with what we've got', it would have been true - we didn't have any apps that required that, but they nevertheless developed those clocks and we made them small and so small that they could put them on a satellite."

Ahead of extra terrestrial GPS one of the first deployments for NPL-CSF3 could be in the more down-to-earth task of helping regulate every other timepiece on the planet. That will impact things from today's generation of GPS and satellites systems to the time-stamping financial transactions in The City and Wall St.

The time standard for all clocks, devices and applications is Universal Co-ordinated Time (UTC), which is maintained by the International Bureau of Weights and Measures (BIPM) outside Paris, France. However, time keepers at the BIPM and International Telecommunications Union (ITU) are coming close to deciding whether UTC should decoupled from astronomical time and come to rely one-hundred-per-cent on International Atomic Time (TAI).

An ITU meeting is scheduled for January in Geneva, Switzerland, where a vote will be taken on decoupling, ending a decade of debate on the topic. Among those objecting to decoupling are Britain's ITU representatives.

See you in 3752, hopefully

UTC is currently set using a network of 300 atomic clocks around the world that are responsible for TAI, but UTC also factors in astronomical time, which is taken from the rotation of the earth. UTC has been used rather than TAI to help mariners as it was felt a system was needed that factored the earth's rotation for accurate navigation on the high seas. Problem is the earth's rotation is not constant, it's slowing down, and every so often a "leap second" is added to ensure accuracy: defined as the Sun crossing the Greenwich meridian at noon UTC to within 0.9 of a second.

Some reckon this means UTC is doomed because the process of adding leap seconds will become unworkable in 1,700 years. Steve Allen, of the Lick Observatory in Santa Cruz, California, said here this summer that based on current calculations UTC will have gone from one leap-second a year in 1981 to one leap second needing to be inserted per month by the year 3752.

This might change as more polar-ice-cap melt, altering the composition of the oceans, slowing the earth further thanks to the dragging effect of water moving around the planet's surface. Alternatively: "If a super volcano erupts or asteroid strikes, all bets are off."

Rise of the time-keeping machines

The BIPM also infavors the death of leap seconds and global convergance on a single, consistant time standard. We are in the process of the earth being surrounded by four satellite systems each with their own separate time offsets, causing confusion and possible disaster the BIPM says. The current US GPS network and Russia's Global Navigation Satellite System (GLONAS) are in the process of being joined by Europe's Galileo system and China's Beidou network.

Elisa Felicitas Arias, director of the BIPM Time Department, told The Reg: "If you have a GPS receiver and are not aware of the differences, you could make a mistake in the order of tenths of a second. If you are landing an aircraft this mistake will have dramatic consequences. We want for such a problem to be avoided if we can do it."

Equally, given so much digital equipment is already installed with in-built offsets it could be argued making any switch also contains dangers.

According to Arias, British delegates have resisted the death of the leap second because no problems have been detected in the past. UTC is, of course, synonymous with Greenwich Mean Time (GMT) so there could be some national pride involved. Disagreement was voiced during a special meeting of experts held under the aegis of the Royal Society in London earlier this month.

If UTC is de-coupled from the earth's rotation, that would mean atomic clocks become completely responsible for setting the time of all other timepieces and apps on Earth.

Unlike UTC, TAI is a constant because its readings are based on a chemical-reaction process that doesn't change. While the reaction is a constant, the process of getting an accurate measurement is a challenge and that challenge arises thanks to the presence external factors that taint the results.

Arias notes, the TAI name could get phased with UTC continuing and the atomic clocks running under the covers of UTC. That could happen because UTC is the legally and internationally accepted system of time measurement while TAI is more of a reference standard. A decision on this could happen at a September 2012 meeting of the ITU following January's vote on decoupling.

In the sidelines of this international struggle is NPL-CSF2. The clock keeps the TAI accurate through its process of exciting caesium atoms using microwaves, and then measuring their reaction. A laser is used to slow down and control the flow of the atoms to help take accurate readings.

The all-important second is measured using a process that takes place when the caesium atoms are exposed to the microwaves: spin flip. The caesium atoms are resonated using microwaves until they flip at a specific frequency – 9, 192,631,770 Hz; a flip defines a second. The atoms are bundled into batches of around 100 million and are streamed through that cavity where they are exposed to microwaves.

Accuracy is not an absolute, however, and NPL is working to make things even more precise.

Looking not unlike a water-tank sitting on polished-steel stilts, the cylindrical NPL-CSF2 lives in controlled, laboratory-like conditions that are free of outside interference. The biggest intrusion in the air-conditioned, windowless room is the quiet tick-ticking of a huge workbench of tiny mechanisms raising and lowering little arms like train track signals to control the lasers used by the clock.

Essen and Parry with Ceasium resonator, pic: Courtesy of NPL

Louis Essen (right) with NPL colleague John Parry and the first caesium clock

Black-body radiation,gravity, and other pesky problems

The caesium atoms inside the clock, though, do get bothered: up to 12 forces from outside create "statistical noise". The same forces affect the other five clocks, too.

These forces include black-body radiation, cold collisions, lasers shining on the wrong atoms, and both the earth's magnetic field and gravity. Take gravity, for example. To measure a second at NPL in London requires a different calibration to the NIST Atomic Clock in Fort Collins, Colorado, to compensate for the effects of gravity. NPL is at an elevation of 79 feet above sea level and suffers more from the effects of gravity on its results than Fort Collins, which is more than 3,315 ft, above sea level in the Rocky Mountains in Boulder Colorado, where the effect of gravity is less.

The hardest factor to eliminate, however, is not gravity: it's that Doppler-like distributed cavity phase the third generation of NPL's clock hopes to mitigate further. Measuring the flip of caesium atoms in the clock is made difficult because the atoms and the microwaves are both moving - hence Doppler effect.

By building a better cavity in NPL-CSF3, the physicists hope they can gain more control over the flow of the atoms for a more precise reading. Gibble's work proposes changes in the geometry of the cavity to reduce further the Doppler effect.

CSF2's cavity has an internal diameter of 4.3cm and a height of 4.3cm, while the component's walls are about one centimetre thick. The external dimensions are larger. Szymaniec reckons NPL will probably approach Gibble to model the new cavity's heat field. "If we are really crazy about reducing this effect we'd follow the design he's proposing," Krzysztof says.

The shape of the new cavity will differ to the one. While still cylindrical, the new cavity will feature staggered end cups to help avoid large phase gradients near sharp edges. Also the microwave energy will be fed to the cavity through four feeds, instead of the current two, to improve the phase distribution uniformity.

NPL-CSF3 isn't the end of work to finally nail the second, however, and NPL, along with NIST in Boulder, the Physikalisch-Technische Bundesanstalt in Germany, SYRTE in Paris and the University of Tokyo are now working on a generation of optical clocks using lasers as the source of radiation rather than microwaves. Laser frequencies are higher so there the errors are smaller and statistical noise is reduced.

"That's in the future; early experiments are promising," Krzysztof told us.

What NPL's man called the "lowest uncertainties" in the results taken from optical clocks have been found in the measurements taken by NIST.

Changing elements

The move to lasers will likely mean the end of the atomic clock as Essen conceived and built it because the optical clocks will need to move off caesium. The reason is that this chemical element doesn't stay excited for long enough using lasers for the physicists to get an accurate reading.

"We'd never have the time to measure the frequency," Krzysztof said.

The optical clocks measure the second either using a process of transitions in single atoms trapped in electromagnetic traps or large numbers of neutral atoms held in traps formed by interfering laser beams – called optical lattices.

The most popular substitutes for caesium so far? Strontium, Ytterbium, and mercury, although NIST uses positively charged aluminium ions – or more precisely AI+ – aluminium with one electron removed to make readings easier.

Szymaniec cautions that while results from optical clocks are encouraging, they are preliminary and such clocks remain in the future. Ahead of that, the third-generation of Essen's offspring is limbering just as humans prepare to depend on it more. ®