Neural network model of two-dimensional silicon-based dielectric cylinder photonic crystal point defect microcavity

Ruichen Xue

doi:10.54254/2977-3903/2025.26451

Advances in Engineering InnovationOpen access

Neural network model of two-dimensional silicon-based dielectric cylinder photonic crystal point defect microcavity

Research Article

Open Access

Neural network model of two-dimensional silicon-based dielectric cylinder photonic crystal point defect microcavity

Ruichen Xue ^1*

¹ Nanjing University of Information Science and Technology

^*Corresponding author: 202283270438@nuist.edu.cn

Published on 29 August 2025

AEI Vol.16 Issue 8

ISSN (Print): 2977-3911

ISSN (Online): 2977-3903

Download Cover

Abstract

Aiming at the problems of high computational cost and long optimization cycle in numerical simulation for traditional photonic crystal microcavity design, this study proposes an intelligent inverse design method combining numerical calculation from the MIT Photonic-Bands (MPB) software package and Back Propagation Neural Network (BP Neural Network). Taking two-dimensional photonic crystals with silicon-based dielectric cylinders (permittivity ε=12) as the research object, 55 sets of band structure data are generated using MPB software by systematically adjusting the radius of dielectric cylinders (0.01–0.5 μm with a step size of 0.005 μm), constructing a "structural parameter-optical property" mapping dataset. A three-layer BP neural network model (with 9 neurons in the hidden layer) is designed, optimized and trained using the Levenberg-Marquardt Algorithm (LM algorithm), combined with the Hyperbolic Tangent Sigmoid Function (tansig function) to handle nonlinear features. Experimental results show that the coefficient of determination R² of the model on the test set reaches 0.95309, with an average relative error of 0.08183% and a maximum relative error of 0.1419%. The design cycle is shortened from the traditional "day-level" to "second-level", with efficiency improved by more than 3 orders of magnitude.

Keywords:

photonic crystals, BP Neural Network, point defect microcavity, inverse design

View PDF

References

[1]. Yablonovitch, E. (1987). Inhibited spontaneous emission in solid-state physics and electronics.Physical Review Letters, 58(20), 2059–2062.

[2]. Painter, O., et al. (1999).Two-Dimensional Photonic Bandgap Defect Mode Laser.Science, 284(5423), 1819-1821.

[3]. Johnson, S. G., & Joannopoulos, J. D. (2001).Block Iterative Frequency-Domain Methods for Maxwell's Equations in a Planewave Basis.Optics Express, 8(3), 173-190.

[4]. Ma, W., Liu, Z., Kudyshev, Z.A.et al.（2021）Deep learning for the design of photonic structures.Nature. Photonics15, 77–90.

[5]. Li, S. Y., Chen, S. W., Jiang, B., et al. (2020). Research progress on neural network-based inverse design of photonic devices.Study on Optical Communications, (3): 33-39.

[6]. Notomi, M. (2000) Theory of Light Propagation in Strongly Modulated Photonic Crystals: Refractionlike Behavior in the Vicinity of the Photonic Band Gap.Physical Review B, 62, 10696-10705.

[7]. Joannopoulos, J. D., et al. (2008). Photonic Crystals: Molding the Flow of Light (2nd ed.). Princeton University Press.

[8]. Ma, D. N., Cheng, H., Tian, J. G., et al. (2022). Inverse design methods and applications of photonics devices (Invited).Acta Photonica Sinica, 51(1), 0151110.

[9]. Hagan, M. T., & Menhaj, M. B. (1994).Training Feedforward Networks with the Marquardt Algorithm.IEEE Transactions on Neural Networks, 5(6), 989-993.

[10]. Yang, S., Wu, Y. Q., Yang, Y., et al. (2022). Neural network-based inverse design method for dual-mode coexisting photonic crystal nanobeam cavity (Patent No. CN114036707A).Chinese Patent.