极化(电化学)
栅栏
光学
材料科学
物理
化学
物理化学
作者
Xin Tu,Yiheng Wu,Tianye Huang,Jing Zhang,Perry Ping Shum
出处
期刊:Journal of The Optical Society of America B-optical Physics
[The Optical Society]
日期:2022-03-14
卷期号:39 (4): 1256-1256
摘要
Polarization-diversity couplers are promising industrially scalable optical devices that can couple optical signals with unknown polarization states into silicon photonic chips. Here we propose a novel, to the best of our knowledge, two-dimensional (2D) metagrating coupler (MGC) for the polarization-diversity system, which improves its coupling efficiency for both polarizations. Compared with the previously reported 2D grating couplers (GCs) with identical grating cells, the proposed apodized design gradually changes the aspect ratio and orientation angles of the elliptical patterns simultaneously. This 2D array of the varied grating cells modulates the diffracted mode field patterns and phases locally, and achieves a better overlap with the Gaussian fiber modes for both polarizations. The peak coupling loss (PCL) for the S-polarized and P-polarized light are − 1.6 d B and − 2.7 d B , respectively. The calculated polarization dependent loss is lower than 0.2 dB from 1522 nm to 1540 nm. The device is robust to the fabrication tolerance and fiber misalignment. The simulated fabrication tolerance analysis show that the etch depth error of ±20 nm results in less than 0.3 dB and 0.1 dB PCL drop, while the grating cell feature size error of ± 40 n m results in less than 0.6 dB and 0.2 dB PCL drop for both types of polarization light, respectively. The fiber misalignment should be within the range from − 2 µ m to + 2 µ m in two perpendicular directions, and the coupling angle deviation should be within the range from − 3 ∘ to + 3 ∘ , in order to assure 0.5 dB penalty loss for both polarizations. The performance of the design is insensitive to the mask misalignments. To assure 0.5 dB penalty loss for both polarizations, Δ θ is suggested to be within the range from − 3 ∘ to + 3 ∘ .
科研通智能强力驱动
Strongly Powered by AbleSci AI