Compact Chip Converts Light into Microwaves for Super-Accurate Timing

Ed Brown

Tech Briefs: How and why did you get started on this project?

Quinlan: Very stable sources of electromagnetic radiation are a long-term interest in the National Institute of Standards and Technology (NIST) Time and Frequency Division. We establish official U.S. time, and part of our mandate is to build next-generation clocks and oscillators. These fill a number of technical and economic needs. Separate from just keeping the trains on time, navigational systems all rely on stable clocks and oscillators. GPS systems are built on putting stable clocks in satellites in orbit — that's how we figure out where we are. Beyond that, radar systems require stable oscillators, a lot of environmental sensing requires stable oscillators, digital-to-analog and analog-to-digital converters all require really, really, stable clocks and oscillators.

So, this has been a long-term interest in the Time and Frequency Division. Then, an opportunity came in the form of a Defense Advanced Research Projects Agency (DARPA) project: to take the tabletop experiments we've been doing in the lab and create a chip-scale version of a very stable microwave oscillator.

Tech Briefs: Why is it that using light to initiate the microwaves makes the oscillator more stable?

Quinlan: The most stable clocks and oscillators are now produced in the optical domain. This is because of the very high frequencies of optical oscillators, where you essentially divide the second into femtoseconds (10⁻¹⁵ seconds). So, you have extremely high precision in the timing. And because optical materials also have extremely low losses, you can circulate the light for such a long time, and it has such a long memory that you keep very, very precise timing, much more so than you can do with microwave systems.

So, we wanted to use an optical system to derive the microwaves, because if we could transfer that exquisite optical stability into the microwave domain, we'd have orders of magnitude more stable radiation than we could otherwise generate.

Tech Briefs: How do you do it?

Quinlan: There are a few key components. First, is the optical reference oscillator, which is our primary timekeeper in the system. Then we need a frequency divider — we need to divide an optical frequency down to a microwave frequency. That is done with technology called an optical frequency comb, which is an amazing tool that's come into physics and engineering in the past 20 or 25 years. It can divide the optical to the microwave. Basically, it sends out a train of optical pulses that are synchronized to the femtosecond oscillator. And then you detect those optical pulses with a high-speed photodetector — an optical to electrical converter: optical pulses in and electrical pulses out.

Those are the key components that we had to reduce to chip scale and get them all working together to make this happen.

Tech Briefs: So, would these components themselves have to be super stable?

Quinlan: You rely on the optical oscillator itself, which is the master oscillator for the system and lock everything to that.

Tech Briefs: Could you describe how the optical oscillator works.

Quinlan: We have a reference cavity that has two mirrors arranged so you can only fit a half integer of wavelengths in the mirrors in order to be resonant. And the reflectivity of the mirrors is 99.999 percent. For contrast, a good metal mirror could be about 99 percent.

Tech Briefs: How do you get them so good?

Quinlan: There's a company here in Boulder, and there are others around the world, that can create these mirrors. They start with a very pristine, very flat glass surface that has angstrom level roughness and then start depositing very thin layers of different refractive index materials to create a stack. Then when light comes in, it all gets reflected back.

Tech Briefs: Wow. That's impressive.

Quinlan: It's an extraordinary technology, I agree.

Tech Briefs: Could you give me some idea of how the frequency comb works.

Quinlan: A mechanical analog would be that it's like a gear set with an optical gear and a huge ratio to get a slower-spinning microwave oscillation at the output.

Tech Briefs: In the article I read, it said that you haven't fit some of the components onto the chip yet. Which are they?

Quinlan: There were separate chips that we connected through optical fiber for this initial demonstration. So, we had the stable lasers on one chip, the frequency comb on another chip, and the detectors on yet another separate chip, all with optical fiber connections in between. Part of this program is to get everything onto one chip, so we are working toward that. Our goal is, at the end of the calendar year, to have everything integrated onto a single chip.

Tech Briefs: How's it going?

Quinlan: At least one of the chips is in the fabrication stage. It's a multi-step process, so the final chip is in the final design stage and some of the intermediate chips are in the fabrication stage.

Also, I should mention that this was a multi-institutional collaborative effort.

Tech Briefs: What do all of these different groups do, and how do they get coordinated?

Quinlan: We have great team management and for our team it's Kerry Vahala at Caltech. In addition to managing the project, he is also producing the chip-scale optical frequency combs.

At the University of Santa Barbara, John Bowers is making ultra-stable lasers.

Peter Rakich’s group at Yale University is making reference cavities.

Steve Bowers and Joe Campbell at the University of Virginia are developing locking electronics and photodetectors.

At JPL, Andrey Matsko and his team are helping with a lot of the design and layout.

Scott Diddams at the University of Colorado is developing a lot of the technology for getting all of the pieces to work in concert, to make sure that the stable lasers and the comb behave well together.

Finally, NIST is doing a lot of the primary evaluation. So, all these parts come to NIST for measurement.

Tech Briefs: How much improvement in the frequency stability do you think you will get?

Quinlan: I think we will be 100 times more stable. We’ve done these types of experiments that fill a lab, but you're not taking a lab with you to do impactful things. So, being able to do that on the chip scale is going to be quite significant.

Tech Briefs: Do you foresee these chips being widely used?

Quinlan: Yes, I do. For all oscillators, working in environments with large temperature swings and that are highly vibrational is extremely challenging. Although we have the same challenges, we have a few advantages by going to the optical realm. It turns out that some of the materials we use are inherently less temperature and vibration sensitive. But packaging it properly is a huge challenge. So that's still going to take a little time.

Tech Briefs: What excites you about this project?

Quinlan: Getting these signals out of the lab — it's getting away from lab demonstrations and into the hands of end users where they can really be applied. I've been working in this field for 15 years and I've done lots of lab experiments and had extensive discussions with other scientists, and now we're finally starting to get it to where it can have real impacts for the larger society.

---