English

Design and analysis of guided modes in photonic waveguides using optical neural network

Optics 2020-08-04 v1 Applied Physics

Abstract

We present a deep learning approach using an optical neural network to predict the fundamental modal indices neffn_{\rm{eff}} in a silicon (Si) channel waveguide. We use three inputs, e.g., two geometric parameters and one material property, and predict the neffn_{\rm{eff}} for the transverse electric and transverse magnetic polarizations. With the least number (i.e., 333^3 or 434^3) of exact mode solutions from Maxwell's equations, we can uncover the solutions which correspond to 10310^3 numerical simulations. Note that this consumes the lowest amount of computational resources. The mean squared errors of the exact and the predicted results are <105<10^{-5}. Moreover, our parameters' ranges are compatible with current photolithography and complementary metal-oxide-semiconductor (CMOS) fabrication technology. We also show the impacts of different transfer functions and neural network layouts on the model's performance. Our approach presents a unique advantage to uncover the guided modes in any photonic waveguides within the least possible numerical simulations.

Keywords

Cite

@article{arxiv.2008.00398,
  title  = {Design and analysis of guided modes in photonic waveguides using optical neural network},
  author = {Nusrat Jahan Anika and Md Borhan Mia},
  journal= {arXiv preprint arXiv:2008.00398},
  year   = {2020}
}