For decades, engineers have faced an uncomfortable trade-off in meeting that demand. Mechanical scanning systems, built around gimbaled mirrors and rotating assemblies, can cover very large angles, but they carry the penalties of size, weight, power consumption, and moving parts that wear out and drift out of alignment. Chip-based optical phased arrays, by contrast, eliminate almost all mechanical motion by electronically controlling the phase of light emitted from many tiny antennas. Yet most phased arrays scan widely in only one direction. Extending them to true two-dimensional operation typically requires densely packed arrays and large numbers of individually controlled elements, which drives up electrical complexity, fabrication difficulty, and energy consumption. The result has been a persistent gap between what free-space optical systems need and what chip-scale technology has been able to deliver.

A research team led by Professor Juejun Hu’s group at the Massachusetts Institute of Technology, working with the Singapore-MIT Alliance for Research and Technology and collaborators, has now demonstrated a hybrid photonic platform that closes much of that gap. Reported in the journal Opto-Electronic Advances, the platform achieves two-dimensional beam steering across a measured field of view of 161 degrees while maintaining nearly diffraction-limited beam quality across a broad angular range. The work was made available online on August 23, 2026, in the Early View section of the journal, and it represents a distinctly different architectural philosophy from the phased-array mainstream.

The key insight behind the new platform comes from an unexpected direction: optical projection. Rather than controlling a dense array of optical antennas with thousands of independently tuned phase shifters, the system uses a silicon photonic integrated circuit as a light-routing engine. The circuit directs light through an on-chip network of switches to a selected output waveguide. At that output, a freeform micro-optical reflector turns the guided light upward, out of the chip plane, and shapes it into a clean free-space beam. In effect, the chip behaves like a projector that selects which of several pre-defined optical paths the light will follow, rather than an antenna array that synthesizes a direction wavefront by wavefront.

The second stage of the architecture is where the wide-angle steering happens. After leaving the reflector, the beam passes through a metasurface—a flat optical element patterned with nanoscale structures that impose precise, position-dependent changes on the light. According to Professor Hu, the position of the selected reflector determines the direction in which the metasurface sends the beam, so choosing a different optical path on the chip directly produces a different steering direction. This division of labor is the heart of the design. The photonic integrated circuit handles the discrete selection of output paths with low complexity, while the metasurface performs the heavy optical work of redirecting and correcting the beam across an ultrawide angular range.

Building the platform required weaving together several fabrication traditions that rarely meet on a single substrate. The team combined analytical optical design, numerical modeling, established semiconductor manufacturing, three-dimensional microprinting, and nanofabrication. The photonic integrated circuit itself was produced using a standard silicon photonics foundry process, meaning the light-routing backbone could in principle be manufactured with the same mature industrial tooling that produces commercial silicon photonic chips. The freeform reflectors, which must efficiently transform the guided optical mode traveling inside a waveguide into a Gaussian-like free-space beam, were designed for that conversion task and then printed directly onto the chip in three dimensions, a technique that allows curved micro-optical geometries impossible with conventional planar lithography.

The metasurface followed a similarly disciplined design pipeline. Its pattern was first established through an analytical design framework, giving the team a physically grounded starting point, and was then refined using optical simulations specifically aimed at reducing aberrations at large steering angles. Aberration control is critical here: when a beam strikes a flat optic far off its nominal axis, simple lenses and gratings tend to smear the focus and distort the wavefront. By optimizing the nanoscale pattern against these off-axis effects, the researchers ensured that beams steered toward the edges of the field of view would emerge nearly as clean as those near the center. After careful assembly and alignment of the chip, reflectors, and metasurface, the complete system was tested at telecommunications wavelengths, the near-infrared band used by fiber networks and many free-space optical links.

The measured results are what make the demonstration notable. The system steered a beam in two dimensions across a 161-degree field of view—an extraordinarily wide cone for any chip-scale device—while keeping the output nearly diffraction-limited, meaning the beam spread was close to the fundamental physical limit set by the aperture rather than degraded by optical imperfections. Because steering is achieved by switching among discrete reflector positions rather than continuously phasing an antenna array, the control electronics remain comparatively simple, avoiding the dense wiring and high power draw that plague two-dimensional phased arrays. The trade-off is that the platform steers among a set of discrete directions, but the researchers see that as a starting point rather than a ceiling.

The architecture, in fact, offers several clear paths toward even better performance. Compact piezoelectric translation could add fine or continuous steering between the discrete directions provided by the current reflector array, blending the simplicity of the switching approach with the smooth coverage of a scanner. Lower-loss photonic foundry processes, improved assembly alignment, and direction-specific metasurface optimization could each raise the overall optical efficiency, putting more of the laser’s power into the steered beam. Increasing the effective aperture, or pairing the platform with beam-expanding optics, would narrow the output beam further, which is essential for long-distance links where divergence over kilometers of free space determines whether enough light arrives at the receiver.

Professor Hu notes that these advances will build on the demonstrated wide field of view and high beam quality, moving the technology toward scalable optical projectors for communications, sensing, and other applications requiring agile, wide-angle control of light. If the hybrid approach matures along those lines, the implications reach across the free-space photonics landscape: inter-satellite optical links that steer across the sky from a chip smaller than a coin, airborne LiDAR with no moving gimbal, and point-to-point Li-Fi communication that can retarget a data beam instantly. A patent based on the technology has been filed by 2Pi Inc., signaling that the researchers see a commercial future for a platform that turns beam steering from a mechanical problem into a manufacturing one.

Subject of Research: Chip-scale hybrid photonic integrated circuit and metasurface platform for ultrawide-angle two-dimensional optical beam steering

Article Title: New chip-scale optical platform enables high-quality ultrawide beam steering

Article References: New chip-scale optical platform enables high-quality ultrawide beam steering. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: beam steering, photonics, metasurface, photonic integrated circuit, silicon photonics, free-space optics, LiDAR, optical communications, micro-optics, nanofabrication, Li-Fi, optical phased arrays

For more than a decade, the dominant paradigm in laser surface engineering has been unidirectional: take a shiny, highly reflective metal and blast it into darkness. By sculpting deep-subwavelength ripples, hierarchical cones, and porous nanostructures into surfaces, researchers have produced remarkable broadband absorbers. In 2025, one team reported a V-scanning strategy that pushed absorption above 99 percent across the 400–700 nanometer window and beyond 98 percent from the ultraviolet all the way to 25 micrometers. Another group achieved uniform blackbodies with absorption over 0.98 in the 3–14 micrometer band using hierarchical cone arrays. But the reverse operation — locally and programmably turning a pre-existing black metal bright again with a laser — had remained largely unexplored, despite its obvious appeal for monolithic device integration.

The research team, led by Mu-Tian Li, Hao Sun, Wei-Wei Xu, and Bing-Rong Gao, began with copper foil chemically etched in a hot alkaline solution containing sodium hydroxide and ammonium persulfate. The treatment grows a dense carpet of copper oxide nanosheets, roughly 30 nanometers thick with lateral dimensions spanning tens to hundreds of nanometers. These randomly arranged, partially overlapping sheets trap light so effectively that the foil loses its metallic luster entirely and appears jet black. It is precisely this metastable oxide architecture that the researchers set out to undo with light.

Their tool was a femtosecond laser system producing 300-femtosecond pulses at a fundamental wavelength of 1030 nanometers, converted to the third harmonic at 343 nanometers through harmonic generation modules. Operating at a fixed repetition rate of 100 kilohertz, the ultraviolet beam was expanded, collimated, steered by a galvanometer scanner, and focused through a 0.5 numerical aperture objective onto the sample in ambient air. The team systematically varied average power from 4 to 40 milliwatts, corresponding to single-pulse energies of 0.04 to 0.40 microjoules and peak fluences of 36 to 362 millijoules per square centimeter, to map out how the surface responds across three distinct regimes.

At low powers of 4 to 12 milliwatts, most nanosheets survived intact; scanning electron microscopy revealed only minor curling, bending, and localized fusion of individual ultrathin sheets, indicating the deposited energy barely exceeded the modification threshold. In the moderate regime of 16 to 28 milliwatts, the transformation became dramatic: the delicate lamellar sheets lost their single-layer character and merged into thicker, plate-like structures roughly 180 nanometers thick, densifying the surface and leaving clearly discernible scan traces. At the highest powers of 32 to 40 milliwatts, overprocessing took over, with vigorous ablation producing irregular debris, recrystallized particles, and highly roughened regions that destroyed most of the original nanosheet architecture.

The processed linewidths, ranging from about 500 to 1200 nanometers, tell an elegant story about Gaussian beam physics. The calculated focal spot diameter is approximately 0.84 micrometers, yet the minimum processed line came in at roughly 500 nanometers — smaller than the spot itself. The explanation lies in thresholding: at low fluence, only the central core of the Gaussian intensity distribution exceeds the modification threshold, so the effective processed width shrinks below the nominal beam diameter. As power rises, a wider annulus of the beam crosses the threshold and the line widens accordingly. Notably, even at elevated power the vertical reach of the process remained limited — the oxide layer was never completely removed, and residual nanolayers persisted within treated zones, a constraint the authors identify as a critical manufacturing limitation: lateral patterning is readily achievable, but vertical depth control is not.

The chemistry behind the brightness change was pinned down by X-ray photoelectron spectroscopy. The pristine black surface was pure divalent CuO, confirmed by Cu 2p peaks at 933.8 and 953.8 electronvolts accompanied by strong shake-up satellites near 942 and 962 electronvolts that arise from the open 3d9 shell of Cu2+. After laser exposure, a pronounced valence transition appeared: low-valence Cu+ states emerged at 932.4 and 951.8 electronvolts, together accounting for 39.72 percent of the copper signal, while the shake-up satellite intensity collapsed to 13.98 percent. The physical driver is photon energy arithmetic — a 343-nanometer photon carries 3.62 electronvolts, comfortably above the roughly 3.04-electronvolt dissociation energy of the CuO bond, so single-photon absorption can cleave the bond and drive oxygen out, first reducing CuO toward Cu2O and, under continued irradiation, potentially toward metallic copper. Energy-dispersive X-ray spectroscopy confirmed the picture, showing copper signals rising and oxygen signals falling along the laser tracks, though roughly 28 percent residual oxygen remained, attributed to incomplete reduction at track peripheries and rapid native re-oxidation in ambient air.

The optical payoff was quantified with an integrating-sphere spectrophotometer across 300 to 2200 nanometers. In the visible band from 350 to 750 nanometers, diffuse reflectance climbed from approximately 4.2 to 4.8 percent on untreated black copper to roughly 7.5 to 8.5 percent after treatment — an approximately 80 percent relative increase based on representative values of 4.5 and 8.0 percent. The authors are careful to note what this does and does not mean: the measurement captures an altered diffuse-reflectance response arising from remodeled morphology and changed composition, but it does not separately quantify enhanced specular reflection versus reduced light trapping, and it cannot by itself predict focusing performance. Under an optical microscope, however, the contrast was unmistakable — laser-written lines and ring patterns glowed visibly brighter against the surrounding darkness.

To prove the concept has device potential, the team direct-wrote amplitude-type reflective Fresnel zone plates onto the black copper, producing clean concentric ring patterns whose circular geometry was well preserved despite slight edge roughness from localized remelting. When observed at different planes along the optical axis, the patterned surface produced a localized bright spot — qualitative evidence of light convergence. The researchers again frame the result cautiously: an ideal binary amplitude Fresnel zone plate would achieve a first-order focusing efficiency of about 10.1 percent under scalar diffraction theory, but finite coherent reflectance, non-ideal amplitude contrast, scattering, and zone-boundary irregularities all conspire to reduce real performance, and no calibrated absolute efficiency was reported. Even so, the demonstration establishes feasibility for in-situ writing of micro-optical components on light-absorbing metallic substrates, and the authors point to clear technological pathways — multi-pulse threshold optimization, processing under inert atmosphere, and auxiliary surface smoothing — for pushing integrated planar micro-optics and laser-based optical encoding toward practical, high-contrast devices.

Subject of Research: Ultrafast laser-induced photoreduction of black copper oxide surfaces for direct-write micro-optics

Article Title: Ultrafast laser-induced photo-reduction on black copper

Article References: Li, M.-T., Sun, H., Chen, Z.-H., Zhuang, R.-J., Huang, W.-W., Hua, J.-G., Shang, P., Xu, W.-W., & Gao, B.-R. (2026). Ultrafast laser-induced photo-reduction on black copper. Results in Optics, Article 101175. https://doi.org/10.1016/j.rio.2026.101175

Image Credits: AI Generated

DOI: 10.1016/j.rio.2026.101175

Keywords: femtosecond laser, black copper, photoreduction, copper oxide, CuO, X-ray photoelectron spectroscopy, Fresnel zone plate, micro-optics, diffuse reflectance, laser surface processing, nanosheets, ultrafast photonics