Laser written waveguides to the sample edge

TL;DR

Edge aberrations near the sample edge reduce focal intensity, with a normalized focal intensity drop to $I/I_0 = 1/4$ when the focus sits directly at the edge. The authors address this with a simple pupil-amplitude mask implemented via the SLM and increased pulse energy to terminate the edge segment and connect it to the bulk, enabling edge writing without polishing. They demonstrate the method on two glasses, FS and E2K, showing that edge-corrected waveguides yield symmetric, well-matched end-face modes and butt-coupling losses below $1$ dB across depths. In the Eagle 2000 substrate, the heating regime helps decouple the laser-induced index change from the focal shape, improving edge-write consistency. This polishing-free edge-writing scheme enables scalable photonic-chip manufacturing and packaging.

Abstract

A method is presented for fabrication of femtosecond laser written waveguides in glass to remove the need for polishing of substrates after processing. It is shown that by amplitude masking the fabrication laser beam near the sample edge and increasing the pulse energy it is possible to write waveguides that are not affected by edge aberrations and display mode profiles well matched to single mode fibre. Results are presented for different depths in fused silica and borosilicate glass substrates. The transmission from fibre to photonic circuit is significantly improved for situations where it is not possible to polish glass substrates after laser writing, creating new opportunities in photonic packaging.

Figures (4)

• Figure 1: (a) Ray tracing showing refraction of the fabrication laser near the edge of the glass substrate, with green rays intersecting the top face of the substrate while red rays are incident on the side face. (b) Reduction of the peak focal intensity as a function of depth $d$ and distance from the sample edge $g$. Results are shown for different NA lenses. Inset shows an example plot of the objective lens pupil plane where only regions in yellow contribute rays that are incident on the top surface of the substrate and included in the focal intensity calculation. (c) Schematic for writing waveguides to the edge. The bulk waveguide is written as normal (left), and intentionally terminated some distance from the substrate edge (centre). The fabrication laser pupil plane is masked to prevent rays intersecting the edge face and the remaining segment of the waveguide to the sample edge is written (right).
• Figure 2: Focal intensity distributions from theory and experiment for different levels of amplitude masking in the pupil plane of the objective lens. Top and middle images show theory for the axial and transverse intensity distributions respectively. Bottom images are experimentally measured, with adjusted intensity to aid visualisation.
• Figure 3: Measured mode field diameter at the sample edge as a function of fabrication depth for laser written waveguides with and without the edge correction technique. Results for E2K (a) and FS (b) glass substrates. Also shown for comparison is the measured mode field diameter for the input optical fiber.
• Figure 4: (a) Example images of the output modes experimentally measured at the output face of the glass chip for waveguides at different depths in E2K and FS, with and without the edge correction technique. (b) Calculated transmission into butt-coupled optical fibre based upon an overlap integral of measured optical modes. Inset is the measured fibre mode. Scale bars are consistent in all images.