Stanford Chip Untangles Mixed Laser Light Into Clean Beams, a Trick Theorized in the 1980s
A fingernail-sized photonic chip sorted partially coherent light from two and then three lasers into separate, uncorrelated beams for the first time.
Stanford engineers and collaborators have built a chip that takes a muddled mixture of light from several lasers and outputs clean, separate beams. It is the first demonstration of a device that accepts partially coherent light and sorts its photons into a set of independent beams, according to a paper in Nature Photonics. The idea has been discussed in theory since the 1980s.
"We're experimentally demonstrating stuff that was thought of originally in the 1980s for the first time," said co-lead author Annie Kroo, a Ph.D. student in electrical engineering at Stanford. "This work has a huge bearing on a whole bunch of different applications because it allows us to think about things in a totally new way."
The physics comes down to coherence. A laser is a coherent source, so its photons oscillate together. Pointed through a mesh, it produces a sharp, speckled pattern because light adds up in some places and cancels in others. A light bulb is incoherent, with photons out of sync, and gives a softer, more diffuse pattern. Most real-world light is partially coherent, a blend of rhythms that line up for a short time and then drift apart.
Co-author Charles Roques-Carmes, an assistant professor at the Institute of Science and Technology Austria and a former Stanford Science Fellow, said the goal is to use light's true nature. "To do this, you have to decompose it into buckets of light that are mutually incoherent, or statistically uncorrelated," he said.
To test that, the team sent light from two infrared lasers into a photonic chip the size of a fingernail. The mixed light was partially coherent. Algorithms and architectures built into the chip sorted it by source and produced separate beams. The researchers then combined those outputs in another device to see if they would add together in intensity. They did not sync up, which showed the beams were mutually completely incoherent. When the team added a third laser, the chip still sorted the mixture into pure beams. "If you look at the outputs, they are completely out of sync; there's absolutely no correlation anymore," said co-lead author Paul-Alexis Mor.
Co-author Shanhui Fan, a Stanford professor, said the chip "learns the underlying coherence modes of an unknown light field and physically directs them into separate outputs." That gives researchers a way to analyze the light's spatial coherence while keeping the separated light available for further processing, said co-author David Miller, Stanford's W. M. Keck professor of electrical engineering, emeritus.
The payoff could be in imaging. A camera records a two-dimensional image, so overlapping sources get lumped together. "In a camera, we throw away the depth information," Miller said. Under a microscope, two overlapping fluorescent particles might look like one, but a chip like this could, in principle, tell the light from the two sources apart.
Miller, co-author Olav Solgaard and Kroo are now working with NASA to see whether the technology could probe light from distant galaxies. Light that looks like a single blob in a telescope might contain separate strands from different stars and exoplanets. "A lot of the information in our world is in light form," Kroo said.
