From Carmon and Vahala, Nature Physics, vol. 3, June 2007, pp. 430–435. © 2007 Nature Publishing Group.

Two adjacent rings can be made to emit different colors, depending on the frequency of the infrared light feeding each one.

Any Color You Like

If you shine a red laser pointer through a glass windowpane you don’t expect it to come out blue on the other side, but with a much brighter beam it just might. At very high intensities light energy tends to combine and redistribute, and red light really can produce blue.

It normally takes brief bursts of megawatts of power to boost light into this high-intensity realm. But now Kerry Vahala (BS ’80, MS ’81, PhD ’85), the Jenkins Professor of Information Science and Technology and professor of applied physics at Caltech, and postdoc Tal Carmon have found a way to do more with less, producing a continuous beam of visible light from an infrared source with less than a milliwatt of power.

At high intensities, light enters the regime of nonlinear optics. We usually notice nonlinearity when there gets to be enough of something to change its environment and rewrite the rules. For example, when a freeway is nearly empty and vehicles effectively have the road to themselves, traffic behaves in a certain way. Put twice as many cars on the road, and the traffic will still behave as if each car owns the road. The only difference is that the flow will double—a proportional, or linear, response. But once traffic nears peak capacity, the vehicles no longer act independently, and the flow becomes miserably nonlinear.

Similarly, light beams pass right through each other at the low intensities we typically encounter, because the photons that make up the beams can usually ignore the cross traffic. At high intensities, however, photons become much more likely to collide and reassemble into other photons—picture three Mini Coopers in dense traffic coalescing into an SUV. The big vehicles of the photon world lie at the higher-energy, or blue, end of the spectrum, with lower-energy photons appearing as red or even infrared light.

From Carmon and Vahala, Nature Physics, vol. 3, June 2007, pp. 430–435. © 2007 Nature Publishing Group.

An end-on view of a beam of blue light coming out of the ring.

Nonlinear optics usually requires brief megawatt intensities, analogous to flooding the freeway with a sudden burst of traffic, but the Caltech researchers attained optical congestion with a much smaller flow by diverting traffic into a tiny no-exit roundabout.

Their traffic circle is a miniscule glass donut, a microresonator smaller across than a human hair. It accumulates power so that a mere milliwatt of infrared light flowing outside the device can sustain an internal flow of 300 watts, a 300,000-fold amplification. Although the infrared light is essentially trapped, energy can still escape as visible light when three infrared photons combine into a single photon of tripled frequency.

Usually researchers in infrared optics can’t directly see their results. This time, Carmon says, “I just turned off the lights and you could see the effect immediately.”

Although infrared light is invisible to human eyes, it is essential to modern telecommunications, flowing through millions of miles of optical fiber. Technology to produce, amplify, and otherwise manipulate near-infrared light is well developed and readily available.

“Our device has several important features,” Vahala says. “First it triples the light frequency, and second, it works in a wide range of frequencies. This means full access to the entire visible spectrum, and likely ultraviolet. Right now there isn’t a way of doing UV generation on a chip. Tunable ultraviolet—that’s exciting.” Coherent UV sources have applications in sensing and also in data storage, where, for example, the laser’s wavelength determines the physical size of the information bit on a compact disk.

The microresonator is part of a promising approach for on-chip optical devices using the silica-on-silicon platform, which is compatible with the electronics of ordinary computer chips. Integrating optics and electronics on the same chip makes the device useful for lab-on-a-chip designs, and the ability to use established fabrication techniques makes large-scale, low-cost production possible.

This work, with Carmon as lead author, appeared in the June 2007 issue of Nature Physics. —JA