| Abstract Scope |
Precision machining of glass is critical for applications in microfluidics, MEMS, and biomedical devices, but tuning laser parameters for clean cuts is difficult due to complex relationships between parameters. We apply Bayesian optimization to efficiently navigate the parameter space and identify ideal cutting conditions. A 4 W, 20 kHz, 260 fs, 1030 nm laser is shaped into a Bessel beam, allowing deeper cuts, using an axicon and focused on 150 µm thick borosilicate glass with a galvanometer and scan lens. Circular cuts, 250 µm in diameter, are scored for quality based on edge roundness, cracking, and debris, which guides optimization. The optimized parameters provide insight into how ultrafast lasers interact with glass, revealing material behavior including ablation thresholds and crack formation. Once optimization plateaus, the best parameters are applied to generate more complex geometries. This approach reduces the time and trial count required to achieve high-quality cuts, offering a scalable solution for advanced glass micromachining. |