Visualization of waveguides fabricated by femtosecond laser micromachining using machine learning assisted limited angle tomography.

We present a machine learning (ML)-assisted limited-angle tomographic approach for visualizing waveguides fabricated by femtosecond laser micromachining (FLM) in glass [1]. Waveguides are inscribed into thin glass via FLM [2] in which an intense femtosecond laser produces plasma and, as plasma undergoes recombination, the index of refraction of the laser-irradiated region increases for light guiding via total internal reflection. Their single-mode guiding performance is validated through light coupling and loss measurements. We employ ML-assisted limited-angle tomography to reconstruct the 3-D refractive index profile of the waveguides, providing detailed structural insight [1]. This approach offers a powerful diagnostic tool for the development and optimization of photonic devices via 3-D visualization.

[1] Jack W. Agnes, Nicole A. Batista, Devi Sapkota, Yuxuan Zhang, Cameron Lapadula, Garima C. Nagar, Dennis Dempsey, Riley S. Freeland, Alexander L. Cuno, James S. Sutherland and Bonggu Shim, Journal of the Optical Society of America B, in press (2025); [2] R. R. Gattass and E. Mazur, Nat. Photonics 2, 219–225 (2008).