Aug. 31, 2004 — A Colorado-based government lab has moved forward in its quest to make a precise clock available to more industries.
Using micro-fabrication techniques, the researchers have created a chip-sized clock, paving the way for its use in handheld devices and other previously impossible applications.
Atomic clocks keep incredibly accurate time, a function vital in virtually every sort of electronics. They work by exciting a small chamber filled with atoms (usually the element cesium), and then measuring the vibrations into one-second intervals. Many atomic clocks can keep perfect time for centuries.
Until now, however, the clocks required large and power-hungry chambers to generate a microwave field, which in turn excited the atoms. They could be used in communication satellites or similar large computer equipment, but not in cellular phones, portable computers, GPS receivers or other rapidly growing fields.
The NIST team solved the problem using a technique called coherent-population trapping, which excites the atoms with light rather than microwaves. That eliminates the need for a large microwave chamber, curbs the power requirement down to that of an AA battery, and allows batch fabrication of the components in glass or silicon wafers.
Coherent-population trapping (CPT) has been known for more than 20 years. The real challenge, NIST physicist and project leader John Kitching said, was to master the fabrication of micro-scale parts to take advantage of the technique.
“People knew CPT would allow much smaller clocks, but that wouldn’t happen without miniaturization techniques,” said John Kitching, a NIST physicist and lead investigator on the project. “That’s where the MEMS comes in.”
Kitching began work on the project five years ago, through funding from the Defense Advanced Research Projects Agency. By the spring of 2001 he had created a prototype about the size of a person’s index finger (atomic clocks are at least a meter tall today). As of last week, the clock measured 1.5 millimeters on each side and 4 millimeters high.
Small-scale atomic clocks are valuable because they allow more devices to operate in synchronized time. Right now a communications system — say, a cellular network — uses a satellite or base-station to broadcast the correct time to handheld units. But that broadcast can be disrupted (by an enemy during military conflict, for example), causing data transmission among various devices to fail. Clocks in the devices themselves would help circumvent such failures.
Kitching is most excited about the batch-manufacturing process, where thousands of clocks could be fabricated on a standard six-inch silicon wafer. Individual components of the clock could be fabricated on separate wafers, which are then stacked atop each other, bonded, and diced into complete, chip-sized atomic clocks. “That opens the door to dramatic cost reductions,” he said.
Still, commercial adoption of the clocks is several years away. Manufacturers must develop MEMS-sized oscillators to measure the atoms’ vibration, along with electronics to control the device and shielding to protect the clock from magnetic fields. The entire clock package will probably be about one cubic centimeter in size.
“I’d put that in the category of engineering rather than science,” said Michael Garvey, vice president of engineering for Symmetricom Corp., a maker of atomic clocks based in San Jose, Calif. “It’s going to take time and money.”
Symmetricom is developing its own atomic clock using light rather than microwaves, with an eye to the GPS market. Kernco Inc., based in Danvers, Mass., developed another small-scale atomic clock with a grant from the U.S. Air Force, but that device is still several inches on each side.
The big fish in the market is Agilent Technologies. Long a pioneer of atomic clocks, the $6.1-billion (in sales) company is working on its own CPT-based clocks, and has kept a close eye on Kitching’s research.
“If the price is right, it could be very useful in many applications,” said Len Cutler, an Agilent physicist and renowned expert on atomic clocks. He estimates that the technology will hit mainstream adoption in another three or four years.