Photonics (Jun 2023)

Design of Silicon Photonics Integrated Bulk Zigzag and Sinusoidal Structured Mode Conversion Devices Using Genetic Algorithm (GA) Optimization

  • Tien-Wei Yu,
  • Chi-Wai Chow,
  • Pin-Cheng Kuo,
  • Yuan-Zeng Lin,
  • Tun-Yao Hung,
  • Yin-He Jian,
  • Chien-Hung Yeh

DOI
https://doi.org/10.3390/photonics10070759
Journal volume & issue
Vol. 10, no. 7
p. 759

Abstract

Read online

To increase the optical interconnect transmission capacity, different multiplexing technologies, including wavelength division multiplexing (WDM), polarization division multiplexing (PolDM) and mode division multiplexing (MDM), can be utilized. Among them, MDM is a promising technique in silicon photonics (SiPh) integrated optical interconnects since higher order modes can be easily generated and preserved in SiPh waveguides. In this work, we propose and demonstrate the designs of SiPh-based bulk zigzag and sinusoidal structured MDM mode conversion devices using genetic algorithm (GA) optimization. A traditional periodic zigzag structured mode converter design has many sharp zigzag angles in the periodic structure, which are very sensitive to the fabrication error. Here, first of all, we propose and demonstrate a bulk zigzag structure to achieve MDM mode conversion. The proposed bulk zigzag structure can reduce the zigzag angle error as a large number of zigzag angles in the periodic structure are eliminated. Moreover, we further improve our device by proposing a bulk sinusoidal structure to further eliminate the zigzag angle. Results show that both the proposed bulk zigzag and sinusoidal MDM mode converters can still maintain high transmissions of >86%, while the mode conversion lengths of both devices can be significantly reduced by >60% in the C-band wavelength window. In addition, as there are many degrees of freedom (DOFs) during the design of the SiPh mode converter, including the waveguide width, length, period, zigzag angle, etch depth, duty cycle, etc., the GA optimization algorithm is employed. Here, detailed implementation of the GA optimization is discussed.

Keywords