Â鶹ÒùÔº

Cornell researchers have built a programmable optical chip that can change the color of light by merging photons, without requiring a new chip for new colors.

This form of nonlinear photonics could potentially be used for classical and quantum communications networks, all-optical signal processing and computation, spectroscopy and sensing.

"Previously, for each combination of colors you wanted to produce, you needed to fabricate a new device with a different design," said Peter McMahon, associate professor of applied and engineering physics in Cornell Engineering, who led the project. "We now have a sort of universal device that lets you do any conversions you want, reprogrammably."

The findings are in Nature. The paper's first author is Ryotatsu Yanagimoto, a former postdoctoral researcher in the McMahon Lab and visiting scientist from NTT Research.

In the regime of linear optics, the frequency—which is to say, the color—of the photons do not change, and the photons tend to ignore each other. Most optics that people encounter in daily life, from eyewear to cell-phone screens, fall into this category.

But in nonlinear optics, the photons interact with each other and can change frequency. Constrained by the laws of energy conservation, one higher-energy photon can be converted into two photons at half the energy. Conversely, two lower-energy photons in a nonlinear optical medium can be combined into one higher-energy photon. McMahon's team performed various demonstrations of the latter with their chip.

To accomplish this, they combined two concepts. The first was to apply a large electric field over the chip via high-voltage probes, enabling frequency conversions in a material that normally wouldn't allow them. The second idea came from a completely different scientific field 20 years ago, when researchers built a device that could manipulate biological cells using a patterned light field to program the device's electric-field distribution.

Co-author Logan Wright, another former postdoctoral researcher in McMahon's group, had realized the same approach could be adapted to make programmable photonic devices.

"By combining these two techniques, we were able to control a material to make it nonlinear in some regions and not nonlinear in other regions," McMahon said. "And it turns out from the theory of nonlinear optics that if you want to control what colors you get out, you need to be able to do this sort of control of the crystal's nonlinearity as a function of space."

The core of the device is a planar-shaped slab of crystal in which light can only travel side-to-side, and not up or down. The researchers sent laser light into this so-called slab waveguide and were able to control how different-colored photons could be combined to produce light with different colors emerging from the chip.

The team made the device at the Cornell NanoScale Science and Technology Facility (CNF) in Duffield Hall, with second author Benjamin Ash '26 playing a prominent role in designing the fabrication process, building devices and testing them.

While the device remains a proof-of-principle demonstration, McMahon said, if the conversion efficiencies were pushed high enough, it could open up a range of opportunities in programmable nonlinear optics, from making new light sources to optical networking.

"Now, in optical networks, people often use different colors of light inside the same piece of fiber to communicate information between computers. Our type of device would be a nice building block to have, on either end of the fiber, letting you reprogrammably change the light's wavelengths," McMahon said.

"There's also a strong argument for it in quantum networks, where we really would like to be able to interface quantum bits that naturally emit photons having different colors. We want to be able to convert that light into the telecom band, as well as back to the natural wavelength of a different quantum bit. Having a device where this is all possible in one package feels like it would be a useful tool."