On a chip no larger than a fingernail, researchers at Stanford and their collaborators have done something physicists had imagined for decades: they took a tangle of partially coherent light, the messy middle ground between a flickering bulb and a disciplined laser, and sorted it into separate beams that no longer interfere with one another. Published at the end of October in Nature Photonics, the work turns a 1980s theoretical idea into hardware that can be held in the hand. Its significance lies less in the device itself than in what it suggests about how much hidden structure ordinary light
Fingernail-Sized Photonic Chip Sorts Tangled Light into Clean, Independent Beams
On a chip no larger than a fingernail, researchers at Stanford and their collaborators have done something physicists had imagined for decades: they took a tangle of partially coherent light, the messy middle ground between a flickering bulb and a disciplined laser, and sorted it into separate beams that no longer interfere with one another. Published at the end of October in Nature Photonics, the work turns a 1980s theoretical idea into hardware that can be held in the hand. Its significance lies less in the device itself than in what it suggests about how much hidden structure ordinary light carries, structure that might one day sharpen microscopes, computers, and our view of distant stars.
Two infrared lasers pointed into a chip about the size of a fingernail, and something researchers had theorized about since the 1980s finally appeared on real hardware. The mixed light going in came out the other side as separate, clean beams. It is the first time a device has taken in partially coherent light and sorted its photons into distinct beams, a result published October 30 in Nature Photonics by a Stanford-led team.
The difference at the heart of the work is one of order. A laser's photons move in step, which is why shining one through a mesh leaves a speckled pattern, while a light bulb's photons drift out of sync and cast a softer grid. Most real light lives somewhere between those extremes, and until now that partial coherence was hard to use. The chip, built on the same principle that moves electrons through laptops and phones but using photons instead, learns the coherence signatures of its inputs and directs each into its own output.
To prove the separated beams were truly independent, the team recombined them and found no build-up of intensity, the signature of beams that never synchronize. They then added a third laser, and the chip sorted that mixture too. Miller, Solgaard, and Kroo now want to know whether the method scales further, and they are working with NASA to see if it can separate the overlapping strands of light from distant galaxies that a telescope currently blends into a single blur.
Two infrared lasers pointed into a chip about the size of a fingernail. That was the core of an experiment at Stanford, and what came out the other side was something researchers had theorized about since the 1980s but had never shown on real hardware: the mixed light going in emerged as separate, clean beams.
The work was done by Stanford engineers and colleagues, and it was published Oct. 30 in Nature Photonics. According to the team, it is the first time a device has taken in mixed, or partially coherent, light and sorted its photons into a distinct set of beams. Annie Kroo, a PhD student in electrical engineering and co-lead author, put it plainly: the group is experimentally demonstrating ideas first conceived in the 1980s, and she said the result bears on a wide range of applications.
Kroo explained the underlying distinction this way. A light bulb is an incoherent source, and a laser is a coherent one. Shine a laser through a mesh and the opposite surface shows a speckled pattern, because the photons oscillate together and in some places add up while in others cancel out. Hold a light bulb to the same mesh and a softer, more diffuse grid appears, since its photons are out of sync with one another.
Most real light sits between those poles. Charles Roques-Carmes, an assistant professor at the Institute of Science and Technology Austria and a former Stanford Science Fellow, said the opportunity lies in exploiting light's true, partially coherent nature for communication, computing, and sensing. To do that, he said, the light has to be decomposed into buckets that are mutually incoherent, meaning statistically uncorrelated.
The chip is a photonic one, which channels photons for data processing much as the microchips in laptops and phones channel electrons. These light-based chips have become smaller and more powerful, though handling the particles inside remains difficult. Olav Solgaard, the Audrey S. Hancock Professor of Engineering at Stanford, said that putting optics on a chip improves them just as putting electronics on a chip does.
In the experiment, two infrared lasers fed the chip, producing a combined light that was partially coherent: the photons' oscillations lined up briefly and then drifted apart. Algorithms and architectures inside the chip sorted that light automatically according to its sources, and it emitted separated beams. To check the result, the researchers sent the beams through a device that recombined them, looking for any addition in intensity. The beams did not sync up, which showed they were independent and completely incoherent with one another. The team then added a third laser, and the chip again sorted the mixture into pure beams. Co-lead author Paul-Alexis Mor, an academic visitor at the Institute of Science and Technology Austria and an engineering student at Corps des Mines in France, said the outputs are completely out of sync, with no correlation left.
David Miller, Stanford's W. M. Keck Professor of Electrical Engineering, Emeritus, said the algorithms identify the unique signatures within the partially coherent light, which lets the chip pull out its coherent components. Co-author Shanhui Fan, the Joseph and Hon Mai Goodman Professor at Stanford, said the chip learns the coherence modes of an unknown light field and physically directs each into its own output, so the light's spatial coherence can be analyzed while the separated light stays available for further processing.
The payoff could be more information than even sophisticated cameras capture. Overlapping sources get lumped together when light is rendered into a two-dimensional image. Miller said a camera throws away depth information, while separating the light yields everything knowable about it. In microscopy, two overlapping fluorescent particles might look like one in conventional imaging, but the chip-based method could in principle tell their light apart.
Miller, Solgaard, and Kroo are now working with NASA to test whether the technology can probe light from distant galaxies, where a single blob in a telescope might hold separate strands from different stars and exoplanets. Whether it scales beyond three lasers and into those settings is the next thing to watch.
Möchten Sie die ganze Geschichte?
Das Original lesen bei Mirage News ↗
Education