Caesium Atomic Clock vs Rubidium and Optical Clocks

Not every atomic clock is the same instrument. A caesium atomic clock in a national laboratory defines the second for the whole world, a rubidium unit no bigger than a paperback keeps a mobile network in step, and an optical clock in a university basement resolves differences neither of them can see. The three families answer different questions, and the differences between them are practical as much as scientific.

This article compares the three by accuracy, size and cost, sets out where each one is used, and explains why optical clocks are expected to redefine the second in the coming decade. To see the end product of all this machinery on your own screen, the live atomic time display reads out server-synchronised time and your device’s drift from it.

How Does a Caesium Atomic Clock Work?

It tunes a microwave signal until caesium-133 atoms absorb it, then counts the cycles of that signal. The SI second is defined as 9,192,631,770 of them, so this design does not approximate the second: it realises the definition directly.

That last point is what sets the family apart. Every other clock is calibrated against something; a primary caesium standard is the something. Two designs are in service. Beam tubes send a stream of hot atoms through a microwave cavity and are compact enough to be rack-mounted, with commercial units holding roughly one part in a million million. Fountain standards such as NIST-F2 in Boulder or NPL-CsF2 in Teddington laser-cool the atoms first, toss them upward and interrogate them for a full second on the way up and down, reaching uncertainties around a hundred thousand times smaller. Fountains are laboratory instruments: they occupy a room, need expert attention and do not travel.

Rubidium Standards: Small, Cheap and Good Enough

Rubidium clocks work on the same principle but use the hyperfine transition of rubidium-87, near 6.835 GHz, inside a sealed glass cell of rubidium vapour. A lamp or laser pumps the atoms into one state and the microwave signal is tuned until absorption peaks. Because everything happens inside a small cell rather than a vacuum tube, the whole standard can fit in a few cubic centimetres, run on a handful of watts and cost a fraction of a caesium unit.

The trade-off is ageing. The gas cell changes slowly over months as the buffer gas and coatings alter, so a rubidium standard drifts in a way a primary standard does not, and it needs periodic discipline from a better reference. That is exactly how it is normally deployed: a rubidium oscillator provides excellent short-term stability and rides through signal outages, while a GPS receiver corrects its long-term rate. Mobile base stations, broadcast transmitters and test equipment are full of them.

Hydrogen Masers

The hydrogen maser sits between the other two in role. It uses the hydrogen hyperfine transition near 1.42 GHz, the same line radio astronomers know as 21 centimetres, and in the active version the atoms themselves oscillate and emit the signal rather than merely absorbing one.

No other device is as stable over minutes to days, which makes masers the workhorse of very long baseline interferometry, deep space tracking and the flywheel clocks of timing laboratories. Over months, though, the cavity and the wall coating shift, so a maser drifts and must be steered by a caesium reference. Passive masers small enough to fly are carried by Galileo navigation satellites.

What Makes Optical Clocks Different?

They divide the second far more finely. A microwave clock counts about nine thousand million cycles per second; an optical clock interrogates a transition at hundreds of terahertz, so a single second contains several hundred million million cycles, and each one is a finer graduation on the ruler.

Two designs dominate. Optical lattice clocks trap thousands of neutral atoms, usually strontium or ytterbium, in a standing wave of laser light tuned so that the trap itself does not shift the transition being measured. Single-ion clocks hold one charged atom in an electromagnetic trap instead, trading signal strength for exquisite control. Both need an optical frequency comb, a laser producing a precisely spaced spectrum of frequencies, to bridge optical frequencies down to countable microwaves; before the comb was developed there was no practical way to turn an optical transition into a clock reading at all. The best of these instruments now reach uncertainties near one part in a billion billion, small enough that raising the apparatus by a centimetre changes its rate measurably through gravitational time dilation.

Caesium Atomic Clock vs Rubidium and Optical

Set side by side, the families separate cleanly on accuracy, size and cost:

• Caesium beam: around one part in a million million, rack-mounted, tens of thousands of pounds, and traceable to the definition of the second without calibration.
• Caesium fountain: a hundred thousand times better again, but a room-sized laboratory instrument that a handful of national institutes operate.
• Rubidium: perhaps a hundred times worse than a beam tube over long periods and subject to ageing, yet small, low-power and cheap enough to embed in ordinary equipment.
• Hydrogen maser: the best short-term stability of any of them, unmatched from minutes to days, but bulky, costly and prone to slow long-term drift.
• Optical lattice: a further hundredfold beyond the fountains, currently experimental, and still hard to keep running continuously.

Notice that accuracy and stability are separate qualities. A maser can be the steadiest device in the building and still sit at slightly the wrong rate, while a caesium standard is right on average yet noisier second by second. Timing laboratories exploit both, letting masers carry the timescale between measurements and caesium tell them where the truth is, a division described in national time standards.

Where Each Family Is Used

Deployment follows the trade-offs. Rubidium goes wherever size, price and power matter more than absolute accuracy: telecommunications, broadcast, instrumentation and the navigation satellites themselves. Caesium beam standards anchor telecom backbones, financial timestamping and military systems that must hold time through a long loss of signal. Fountains and masers stay in the national laboratories, where NIST, NPL, PTB and their counterparts compare them and send the results to the BIPM to be combined into International Atomic Time. Optical clocks are mostly compared against each other for now, and increasingly used for physics that has nothing to do with timekeeping, such as measuring height differences by their effect on the flow of time. The satellite case is covered in atomic clocks in GPS satellites.

Why Optical Clocks Will Redefine the Second

The definition of the second has always followed the best available clock, and optical standards now outperform the caesium fountains that realise it. Several optical transitions in strontium, ytterbium and trapped ions are already recognised internationally as secondary representations of the second, and metrology institutes are working toward a redefinition around the 2030s.

The remaining obstacles are engineering rather than physics. A defining standard has to run continuously and be comparable between continents, and comparing optical clocks over long distances is harder than comparing microwave ones, since satellite links are not yet precise enough and optical fibre links must be built. When it does happen, nothing on your wall will change by a visible amount. The new definition will be tied to the old one so that the length of a second is preserved, exactly as it was in 1967, and the scales you can read on the TAI and GPS time page will continue uninterrupted.

Conclusion

A caesium atomic clock defines the second, a rubidium unit makes atomic timekeeping cheap and portable, a hydrogen maser supplies unmatched short-term steadiness, and optical lattice clocks point at the standard the world will adopt next. Choosing between them is not a matter of picking the best number but of matching accuracy, size, cost and endurance to the job. The mechanism common to all of them is set out in how atomic clocks work. To see the result rather than the hardware, open the atomic time display or begin at the atomicclock.now homepage.