The authors investigate whether direct high-resolution 3D printing can realize short, geometrically complex photonic crystal fiber waveguides and integrate them on optical fiber tips. They fabricate and optically characterize multiple waveguide designs and a 210 µm polymer polarization beam splitter composed of a down-taper, dual-core directional coupler, and fan-out. Iterative simulation, fabrication, and measurement produce a coupling-length ratio of 1.97; the fiber-mounted splitter experimentally exceeds 10 dB extinction over a 100 nm bandwidth around 1550 nm.

A 3D-printed structure on the end of an optical fiber sorts light by the direction it vibrates.

Light has a direction you can’t see.

As light travels, its electric field vibrates across its path. That orientation is called polarization. The researchers built a microscopic polymer device that sends two perpendicular polarizations into separate guiding cores. An array of microscopic air channels shapes how light travels in the solid polymer cores. The trick is to let light move between two nearby cores—and stop at just the right moment.

Turn the light. Change the split.

Rotate a linearly polarized input. See how much light leaves through each output of an ideal splitter.

Control the incoming polarization

Schematic · colors identify perpendicular polarizations, not wavelengths

An ideal, lossless splitter routes cos²θ of the input power to A and sin²θ to B. The real device has losses and imperfect separation.

Two rhythms. One perfect exit.

Nearby cores exchange light. Their geometry makes the two polarizations swap power at different rates.

Find the splitting length

Vertical axis: fraction of each polarization transferred to the other core

A clean split: coral has transferred to the other core; cyan has returned to its starting core.

A tiny, three-part journey.

Two-photon lithography prints the polymer structure directly onto a conventional optical fiber.

Let the cores talk.

In the dual-core region, light spreads enough to reach the neighboring core. Different coupling rates let one polarization transfer while the other returns.

Small device. Measurable separation.

The researchers tested the printed splitter with near-infrared light around 1,550 nm.

Coupling-length ratio

The achieved ratio sits close to the ideal 2 : 1 required for the two exchange rhythms.

Measured extinction over 100 nm

Extinction compares wanted and unwanted polarization power in an output. 10 dB means a 10-to-1 ratio.

Minimum extinction at 1,550 nm

Both outputs exceeded 12.6 dB across the 1,530–1,565 nm telecommunications C-band.

Real light is not lossless.

Measured insertion losses at 1,550 nm were 1.18 dB for horizontal and 1.35 dB for vertical polarization. Polymer absorption, residual scattering and higher-order modes limit performance. Fabrication shrinkage also matters: simulations alone did not predict the final behavior reliably, so experimental tuning was essential.

3D printed waveguides based on Photonic Crystal Fiber designs for complex fiber-end photonic devices ↗

Andrea Bertoncini & Carlo Liberale · arXiv:2009.04209v1 · 7 September 2020

This experience focuses on the paper’s polarization beam splitter. Interactive models are idealized teaching tools; performance figures above are reported measurements. Mechanism: Section 2 and Figure 2; measurements: Section 2 and Figure 5; coupling model: Supplement S6. Hero artwork is an AI-generated conceptual illustration.