MIT researchers have built a nanoscale photonic device shaped like a tiny ski jump that launches laser light directly off a silicon chip and steers it through free space. Reported by MIT on March 11, 2026, and described in a Nature paper and companion preprint, the work tackles a longstanding bottleneck in integrated photonics: getting light off a chip efficiently without bulky mirrors or lenses. The device could help simplify how quantum systems, compact displays, and augmented-reality hardware deliver and steer light.
How a Waveguide Becomes a Ski Jump
The core innovation is deceptively simple in concept but demanding in execution. The team fabricated a silicon nitride (Si3N4) waveguide on top of a piezoelectric aluminum nitride (AlN) microcantilever. Because the two materials expand at different rates, the cantilever passively curls upward once released from the chip surface, bending roughly 90 degrees out of plane. The result is a ramp-like structure that guides photons along its length and then fires them upward as a tightly focused beam.
That passive curl is the key differentiator. Traditional on-chip light emitters, such as grating couplers, scatter light at shallow angles and lose energy in the process. The ski-jump geometry instead points the beam nearly straight up, and applying voltage to the piezoelectric layer lets operators steer the emitted beam across a range of angles. In testing, the device achieved a scanning figure of merit of 68 million spots·s⁻¹·mm⁻², a metric that captures both speed and spatial density of addressable points, according to the team’s expanded preprint.
Why Getting Light Off-Chip Has Been So Hard
Photonic integrated circuits can route light with extraordinary precision on a flat chip. The trouble starts at the edge. Coupling that on-chip light into free space, or into an optical fiber, typically requires external optics that are orders of magnitude larger than the chip itself. A 2017 review by M. J. R. Heck catalogued the state of optical beam scanning and found that most approaches still relied on mechanical components, such as rotating mirrors or galvanometers, that resist miniaturization.
Grating couplers offered one alternative, and MIT Lincoln Laboratory has explored them for trapped-ion quantum computing, where lasers must hit individual ions with micrometer precision. But gratings radiate light over a broad angular cone, wasting photons. Optical phased arrays, another on-chip strategy, can steer beams electronically but grow complex and power-hungry as the number of emitters scales up. The ski-jump sidesteps both problems by physically redirecting the waveguide itself, turning the emission direction from in-plane to out-of-plane without splitting the beam or relying on interference patterns.
Quantum Applications and Diamond Color Centers
The most striking near-term application sits at the intersection of photonics and quantum science. The researchers demonstrated that their ski jump could be used to excite diamond colour centres, which are atomic-scale defects in diamond’s crystal lattice that emit single photons on demand. Diamond colour centres are among the leading candidates for quantum networking nodes and quantum sensors, but addressing them usually requires free-space optics mounted above the sample, a setup that does not scale well.
If a ski-jump emitter can target individual colour centres from a chip below, it opens a path toward fully integrated quantum photonic modules where the light source, the steering mechanism, and the quantum emitter all sit within a few hundred micrometers of each other. Off-chip optical loss is a dominant performance bottleneck in quantum experiments, as recent work on on-chip phased arrays for non-classical light has shown. Eliminating even a fraction of that loss by removing bulk optics could meaningfully improve entanglement rates and signal fidelity in compact quantum devices.
Beyond Quantum: Displays, Lidar, and Beam Multiplexing
Quantum systems are not the only beneficiaries. MIT described the ski jump as part of a new class of photonic devices that enable precise broadcasting of light, with potential uses in augmented-reality glasses and compact displays. Lidar sensors for autonomous vehicles face a similar design tension: they need fast, wide-angle beam scanning from a package small enough to fit behind a car’s bumper. A steerable emitter that operates at 68 million addressable spots per second per square millimeter, if it can be manufactured at scale, could offer an alternative to current microelectromechanical (MEMS) mirror arrays in speed and footprint.
The broader photonics community has been chasing chip-to-free-space beam control from multiple directions. In June 2023, NIST announced a photonic chip that transforms a single beam into multiple beams with varied properties, a different approach focused on beam multiplexing rather than beam steering. Together, these efforts suggest that the field is converging on a future where chips handle not just light generation and routing but also the final mile of shaping and directing photons into the outside world.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
