What is an atomic clock?

An atomic clock is a device that produces electronic 'ticks' at a rate related to certain electromagnetic oscillations of atoms. Atoms produce some of the most stable periodic signals currently known. Therefore, a clock based on these oscillations can be extremely precise. One might compare an atomic clock to a grandfather clock. At the heart of a grandfather clock is a pendulum, which swings back and forth in a regular, periodic way. Every time the pendulum reaches the middle of its swing, a 'tick' is produced (by the mechanics behind the face of the clock) that moves the second hand of the clock forward by one second. The more stable the rate of swinging of the pendulum, the more precise the clock will be. In an atomic clock, the atoms are the equivalent of the pendulum. The clock is designed so that every time the oscillating components of the atoms reaches a certain point in their periodic motion, an electronic 'tick' is generated, which can be used to run an electronic timer or counter.

How does a passive atomic clock work?

Imagine pushing a child on a swing. When pushed once, the swing, like the pendulum in a grandfather clock, has a specific time that it takes to return to its starting point. This time is called its 'period'. If you want the child to swing high, you need give the swing regular pushes separated in time by exactly the period of the swing, for example only when the swing has reached the peak of its backward motion. If, on the other hand, your swing pushes are not separated in time by the period of the swing, then soon you'll be pushing at the wrong time and slowing the swing down rather than speeding it up, and the maximum height reached by the swing will be smaller. You would therefore be able to tell whether the period of your pushing corresponded to the period of the swing by measuring the height the swing reaches after you've been pushing for a while.

In a passive atomic clock, atoms are excited (pushed) by an oscillating electromagnetic field  (you), similar to the one in a microwave oven. The excitation field is produced by a part of the clock often referred to as the 'local oscillator'. When the period of the excitation field exactly corresponds to the period of oscillation of the atoms, the atomic oscillation (swing motion) reaches a large amplitude. The amplitude of the atomic oscillation can be measured, for example by monitoring the absorption of the power of a light beam traversing the atoms (volume of child laughing). If the period of the excitation field does not correspond to the period of the atomic oscillation, the atomic oscillation will be small, changing the absorption of the light beam (child silent). The part of the atomic clock that contains the atoms and compares the period of the local oscillator's electromagnetic field to the period of the atomic oscillation is known as the 'physics package' (swing plus child). The 'control system' of the clock (your brain) takes the output of the physics package, an electronic signal proportional to the amplitude of the atomic oscillation, and corrects the period of the local oscillator's field to make it exactly equal to the period of the atoms. In this way, the local oscillator attains the intrinsic stability of the atoms and can be used to generate the ticks that are the output of the clock.

What is coherent population trapping and how is it used in chip-scale clocks?

In most atomic frequency references, the stable atomic oscillation is excited by applying a microwave field directly to the atoms. Typically, the cell containing the atoms is placed inside a microwave cavity, which confines the microwave field to the vicinity of the atoms. The technique of coherent population trapping (CPT) was developed in the 1970s as an alternate way to excite microwave resonances in atoms. Instead of applying a microwave field to the atoms, two light fields are applied which themselves differ in frequency by the microwave frequency. These two optical fields can excite a similar resonance in the atoms as in the direct microwave case. CPT can be implemented by modulating the injection current of a diode laser and using the resulting sidebands on the optical carrier to excite the resonance. CPT excitation allows a considerably simplified frequency reference design and possible advantages with regard to overall size and power dissipation. CPT excitation of narrow microwave resonances has been demonstrated at NIST in millimeter-scale cells.

Why do we want to miniaturize atomic clocks?

Atomic clocks provide a critical backbone for a variety of technological systems. For example, the global positioning system (GPS) is based on a collection of atomic clocks orbiting the earth in satellites. As another example, many digital communication systems, such as cellular telephone networks rely on precision timing from atomic clocks to synchronize the network and enable efficient, error-free transfer of information. In many cases the stability of the clock translates directly into how well the system it supports can perform. The positioning accuracy of GPS is closely tied to the stability of the clocks in the GPS satellites. High-stability clocks, such as atomic clocks, are therefore very valuable. However, there are many potential applications that are not able to benefit from the stability of current generation of atomic clocks because the clocks are too big, or consume too much power. For example, current atomic clocks cannot be installed in cell phones themselves, GPS receivers or pretty much any other battery-operated, portable electronic device. The performance of some of these devices could benefit substantially from highly miniaturized, low-power clocks that retain some of the high stability typical of atoms. Click here for more information on applications.

How stable are NIST's chip-scale atomic clocks?

The chip-scale atomic clocks being made at NIST are stable to better than one part in ten billion when timing events over one second. This is equivalent to neither gaining nor loosing one second over 300 years. By comparison quartz crystal oscillators, often found in wristwatches, are stable to about 1 second over a few days. High-performance laboratory atomic clocks are stable to one second over 100 million years.

How small are the NIST chip-scale atomic clocks?

The physics packages being built at NIST measure 1.5 mm  x 1.5 mm x 4.2 mm (1 mm is 40 thousands of an inch), or about the size of a grain of rice. The cell that contains the atoms is about the size of a grain of sand and contains about one billion cesium atoms in vapor form.

Is there anything unique about the way these chip-scale atomic clocks are made?

Yes! All atomic clocks until now have been built by fabricating individual components (atomic cells, lamp, control circuit board, etc.) and assembling them one-by-one into complete units. This is a costly, time-consuming process and also leads to significant variation from unit to unit in the timing output that must be corrected after assembly. The fabrication process designed at NIST allows many components of the physics package to be made at one time on large (6 inch) wafers. Since it's essentially just as easy to make an entire wafer of components as it is to make one, this advance allows for a huge savings in fabrication cost. In addition, once a wafer of each component has been made, the clocks can all be assembled together by just stacking the wafers, bonding them together and then dicing the stacked structure into individual components. We expect that several thousand individual clock physics packages could be made with one single process sequence.

This advance in the fabrication and assembly method for atomic clocks is similar to the advance that took the field of electronics from circuits assembled from discrete components to integrated circuits. This change in electronics manufacturing procedure heralded the computer age since it allowed large numbers of electronic components, such as transistors, to be integrated in high densities on a single computer chip at low cost. It is possible that microfabricated atomic clocks could also be integrated, in an efficient, low-cost manner, into other electronic devices to provide stable timing signals.

Has NIST built all parts of these new atomic clocks?

No. NIST has so far just developed the physics package, which is the heart of the clock and contains the atoms that provide the stable frequency reference. Although it is likely that this will end up being the most difficult and important step in the development of microfabricated commercial frequency references, other challenges remain. For example local oscillators compatible with the goals of the program are still under development and will need to be integrated with the NIST physics package, as well as some appropriate miniaturized control electronics, before a complete microfabricated atomic clock can truly be said to have been demonstrated. It is expected that with further research, a complete atomic clock could be constructed that has a volume of 1 cm³, and dissipates less than 30 mW of electrical power.

Are there any other useful devices that could be made with this technology?

Most certainly. Atoms are also excellent sensors! Almost any environmental perturbation, such as a change in ambient temperature, pressure and local electric and magnetic field, changes the rate of the 'ticks' produced by the atom. By comparing two atomic systems, one of which has been designed to be maximally sensitive to the perturbation (i.e., configured as a sensor) and the other of which has been designed to be minimally sensitive (i.e., configured as a clock), it is possible to measure the strength of the perturbation. For example, magnetic fields can be measured in this way, and microfabricated magnetic field sensors based on atoms are presently under development at NIST. These miniature devices could have field sensitivities orders of magnitude beyond what is possible with other technologies of equivalent size.

What kind of laser is used in the chip-scale atomic clock?

The type of laser used is a vertical-cavity surface-emitting laser, or VCSEL (pronounced 'vicksel'). These lasers were developed for the telecommunications industry over the last fifteen years, can run on very low power and are highly reliable. In contrast to edge-emitting lasers, light from a VCSEL is emitted vertically, through the top of the device. This makes the vertical integration of the other components such as the optics assembly and the cell easier because the light is already traveling in the vertical direction.

Return to Microclock Home