Photonics: Ski jump structure; on-chip nanolasers; neural networks.
Researchers from Massachusetts Institute of Technology (MIT), MITRE, University of Arizona, and Sandia National Laboratories developed a new class of photonic devices that enable the precise broadcasting of light from a chip into free space.
The chip uses an array of microscopic structures that curl upward, resembling tiny ski jumps, and allows control over how light is emitted from thousands of the structures at once. Applications include both high-resolution compact displays and photonics-based quantum computing.
The chips are constructed using a two-layer structure of patterned silicon nitride and aluminum nitride, which expand at different rates when cooled after manufacturing. The difference in strain between the layers causes the structure to curve upward.
“On a chip, light travels in wires, but in our normal, free-space world, light travels wherever it wants. Interfacing between these two worlds has long been a challenge. But now, with this new platform, we can create thousands of individually controllable laser beams that can interact with the world outside the chip in a single shot,” said Henry Wen, a visiting research scientist in the Research Laboratory of Electronics at MIT and research scientist at MITRE, in a press release. “Both of these materials, silicon nitride and aluminum nitride, were separate technologies. Finding a way to put them together was really the fabrication innovation that enables the ski jumps.”
Connected waveguides on the chip funnel light to the ski jump structures. A series of modulators are used to control how that light is turned on and off, enabling the researchers to project light off the chip and move it around in free space.
Next, the team plans to scale up the system to conduct experiments on yield, uniformity, and robustness. “We envision this opening the door to a new class of lab-on-chip capabilities and lithographically defined micro-opto-robotic agents,” Wen added. [1]
Researchers from the Technical University of Denmark developed an on-chip nanolaser based on a light-trapping nanocavity in a semiconductor membrane. When the laser is illuminated with a beam of light, both light and electrons are concentrated in a very small area, enabling the laser to operate at room temperature with low energy consumption.
“The nanolaser opens up the possibility of creating a new generation of components that combine high performance with minimal size. This could be in information technology, for example, where ultra-small and energy-efficient lasers can reduce energy consumption in computers, or in the development of sensors for the healthcare sector, where the nanolaser’s extreme light concentration can deliver high-resolution images and ultrasensitive biosensors,” said Jesper Mørk, a professor in the Department of Electrical and Photonics Engineering at DTU, in a statement.
Next, the researchers plan to investigate powering the laser electrically to make it feasible for photonic chips and sensors. They believe the technical challenges can be solved within 5-10 years. [2]
Researchers from the University of Sydney designed a photonic AI chip prototype that implements neural network models directly in the photonic structure. The inverse-designed photonic neural network accelerator was trained to classify biomedical images, such as breast, chest, and abdomen MRI scans, and achieved between 90-99% classification accuracy.
“We’ve re-imagined how the photonics can be used to design new energy efficient and ultrafast computer processing chips,” said Xiaoke Yi, a professor from the School of Electrical and Computer Engineering and director of the Photonics Research Group at the University of Sydney, in a statement. “Artificial intelligence is increasingly constrained by the energy consumption. This research performs neural computation using light, enabling faster, more energy-efficient and ultra-compact AI accelerators.”
Next, the team plans to create larger-scale photonic neural networks. [3]
[1] M. Saha, Y.H. Wen, A.S. Greenspon, et al. Nanophotonic waveguide chip-to-world beam scanning. Nature 651, 356–363 (2026). https://doi.org/10.1038/s41586-025-10038-6
[2] M. Xiong, Y. Yu, Y. Berdnikov, et al. A nanolaser with extreme dielectric confinement. Sci. Adv. 11, eadx3865 (2025). https://doi.org/10.1126/sciadv.adx3865
[3] J. Sved, S. Song, L. Li, et al. Inverse-designed nanophotonic neural network accelerators for ultra-compact optical computing. Nat Commun 17, 1059 (2026). https://doi.org/10.1038/s41467-026-68648-1
 |
 |
 |
 |
 |  |
 |  |
Leave a Reply