3D-Printed Photonic Lanterns Merge 37 Lasers Into a Single Fiber in Under Half a Millimeter
The Hebrew University of Jerusalem
Combining light from many small lasers into a single powerful beam is conceptually simple and practically difficult. The physics of beam combining typically requires bulky optical systems with precise alignment of lenses, mirrors, and fibers. A team at the Hebrew University of Jerusalem has shrunk that entire apparatus into a 3D-printed structure smaller than the period at the end of this sentence.
The VCSEL challenge
Vertical-cavity surface-emitting lasers (VCSELs) are tiny semiconductor lasers that emit light perpendicular to their surface. They are cheap to manufacture in large arrays, efficient, and widely used in applications from data communications to facial recognition sensors. But each VCSEL in an array emits its own multimode beam, meaning the light contains multiple spatial patterns bouncing around inside the laser cavity. Getting all that light from dozens of VCSELs into a single optical fiber efficiently has been a persistent engineering bottleneck.
Traditional relay lens systems can combine beams, but they are physically large, require precise alignment, and often sacrifice beam quality in the process. The brightness of the combined beam, a critical measure of optical quality, typically degrades as more sources are added.
The first multimode photonic lantern
Photonic lanterns are optical devices that transition light between multiple waveguides and a single multimode waveguide. Until now, they were designed exclusively for single-mode inputs, making them incompatible with the multimode output of VCSEL arrays.
The Hebrew University team, led by PhD student Yoav Dana under the guidance of Professor Dan Marom, solved this by designing what they call an "N-MM PL" architecture: a photonic lantern that accepts many multimode inputs and merges them into a single high-mode-count output waveguide. The key insight was matching the total number of spatial modes across all inputs to the modal capacity of the output fiber, preserving brightness through what is essentially a conservation of optical information.
The devices are fabricated using two-photon polymerization, a 3D printing technique that builds structures with sub-micrometer resolution by selectively hardening photosensitive resin with a focused laser.
Performance across three scales
The team demonstrated photonic lanterns at three scales: 7-input, 19-input, and 37-input configurations, with each input VCSEL source lasing across six spatial modes. The 37-input device supports a total of 222 spatial modes in a single output fiber.
Coupling losses into standard 50-micrometer multimode fiber were remarkably low: -0.6 dB for the 19-input device and -0.8 dB for the 37-input device. In practical terms, the 19-input lantern transmits about 87% of the incoming light, and the 37-input version transmits about 83%. For an incoherent beam combining device at this scale, those numbers are exceptional.
The physical dimensions are striking. The 37-input photonic lantern measures just 470 micrometers in length, less than half a millimeter. Traditional optical multiplexing systems that achieve comparable channel counts are orders of magnitude larger.
Why brightness preservation matters
In many applications, raw optical power is not enough; brightness, the power per unit area per unit solid angle, determines how effectively light can be focused, transmitted through fiber, or delivered to a target. Traditional beam combining methods that simply merge beams into a larger fiber increase total power but dilute brightness by spreading light across a larger modal volume.
The N-MM PL approach avoids this by carefully matching the degrees of freedom. Each input VCSEL contributes a defined number of spatial modes, and the output waveguide is designed to support exactly the sum total of all input modes. No modal capacity is wasted, and no brightness is sacrificed to unnecessary mode volume.
Applications and limitations
The technology targets applications where efficiently delivering high optical power through a single fiber matters: high-power laser systems, optical communications requiring high aggregate data rates from VCSEL arrays, medical laser systems, and industrial processes. The collaboration with Civan Lasers, an Israeli laser company, and funding from the Israel Innovation Authority signal commercial interest in translating the research into products.
Several limitations should be noted. The devices have been demonstrated in laboratory conditions with controlled alignment and environmental stability. How they perform under the vibration, temperature variation, and mechanical stress of real-world deployment is not yet characterized. The two-photon polymerization fabrication process, while capable of extraordinary resolution, is slow compared to mass-manufacturing techniques. Scaling production to commercial volumes will require either process acceleration or alternative fabrication methods.
The polymer material used in 3D printing also has limitations in optical power handling. At very high power levels, the material may degrade or exhibit nonlinear effects. Whether these photonic lanterns can handle the power levels required for industrial or defense applications has not been demonstrated.
The approach is specific to incoherent beam combining, where the phase relationships between sources do not matter. Applications requiring coherent combination, where all beams must be in phase to produce a diffraction-limited output, need different techniques.
Shrinking the bottleneck
The progression from 7 to 19 to 37 inputs suggests the architecture has room to scale further, though each step up increases fabrication complexity and tightens alignment tolerances. If the manufacturing challenges can be addressed, the approach could substantially simplify optical systems that currently require complex free-space optics, replacing racks of lenses and mirrors with a printed polymer structure smaller than a grain of sand.