Lu Yilin, Chen Ruohui, Hu Chao, Liu Xiaofei, Gan Zhikai, Zhao Xingyan, Qiu Yang, Zheng Shaonan, Zhong Qize, Dong Yuan, Lin Chun, Hu Ting
Appl Opt. 2024 Apr 20;63(12):3242-3249. doi: 10.1364/AO.520895.
Traditional long-wave infrared polarimetry usually relies on complex optical setups, making it challenging to meet the increasing demand for system miniaturization. To address this problem, we design an all-silicon broadband achromatic polarization-multiplexing metalens (BAPM) operating at the wavelength range of 9-12 µm. A machine-learning-based design method is developed to replace the tedious and computationally intensive simulation of a large number of meta-atoms. The results indicate that the coefficients of variation in focal length of the BAPM are 3.95% and 3.71%, and the average focusing efficiencies are 41.3% and 40.5% under broadband light incidence with - and -polarizations, respectively.
传统的长波红外偏振测量通常依赖于复杂的光学装置,这使得满足对系统小型化日益增长的需求具有挑战性。为了解决这个问题,我们设计了一种全硅宽带消色差偏振复用超透镜(BAPM),其工作波长范围为9 - 12微米。开发了一种基于机器学习的设计方法,以取代对大量超原子进行的繁琐且计算量大的模拟。结果表明,在宽带光以 - 偏振和 - 偏振入射时,BAPM的焦距变化系数分别为3.95%和3.71%,平均聚焦效率分别为41.3%和40.5%。