期刊
OPTICS LETTERS
卷 46, 期 17, 页码 4422-4425出版社
Optica Publishing Group
DOI: 10.1364/OL.436118
关键词
-
类别
资金
- New Energy and Industrial Technology Development Organization [JPNP13004]
Machine learning was used to model the photonic bands of SiO2-dadded Si lattice-shifted photonic crystal waveguides, leading to the discovery of a structure that can generate low-dispersion slow light with a group index of approximately 20 in the full C-band at telecom wavelengths. The transition structure between this waveguide and a Si-channel waveguide was designed using an evolutional optimization, resulting in a C-band average loss of 0.116 dB/transition. These results suggest the potential for further enhancing the versatility of slow light.
We modeled the photonic bands of SiO2-dadded Si lattice-shifted photonic crystal waveguides via machine learning and found a structure that generates low-dispersion slow light with a group index of approximately 20 in the full C-band at telecom wavelengths. The normalized delay-bandwidth product is as large as 0.45, which is close to the theoretical upper limit. The transition structure between this waveguide and a Si-channel waveguide was designed using an evolutional optimization, and a C-band average loss of 0.116 dB/transition was calculated. These results prove the possibility of further enhancing the versatility of slow light. (C) 2021 Optical Society of America
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据