【摘要】 设计出一种由氮化钛（TiN）衬底、二氧化硅（SiO₂）电介质层和TiN图案层构成的紫外线超材料吸收器。采用有限元法（FEM）分析该吸收器的吸收特性并建立数据集。采用基于串联神经网络的深度学习模型对该吸收器进行逆向优化设计，极大地加快了设计过程。结果表明：采用优化后的结构参数，吸收器在200～400 nm范围内，在正入射条件下，吸收率大于94.3%，其中265～375 nm范围内的吸收率可达到97.0%；在50°入射条件下，横磁（TM）波入射后的平均吸收率达到82.6%，横电（TE）波入射后的平均吸收率达到83.0%。该吸收器的吸收机理为腔共振效应。与现有文献报道的吸收器相比，所设计的紫外线超材料吸收器不仅结构简单、易于实现，还在紫外波段展现出更高的吸收率，且具有大入射角稳定性及偏振无关性。更多还原

【Abstract】 An ultraviolet metamaterial absorber is designed in this paper, consisting of a titanium nitride（TiN） substrate, a silicon dioxide（SiO₂） dielectric layer, and a TiN patterned layer. The absorption characteristics of the absorber are analyzed using the finite element method（FEM）, and a dataset is established. A deep learning model based on a tandem neural network is employed for the inverse optimization design of the absorber, significantly accelerating the design process. It is demonstrated that with the optimized structural parameters, an absorptivity greater than 94.3% is achieved in the 200-400 nm range under normal incidence, with absorptivity of 97.0% in the 265-375 nm range. Under a 50° angle of incidence, an average absorptivity of 82.6% is obtained for transverse magnetic（TM） wave incidence, while an average absorptivity of 83.0% is observed for transverse electric（TE） wave incidence. The absorption mechanism of the absorber is attributed to cavity resonance effects. When compared to absorbers reported in existing literature, the ultraviolet metamaterial absorber proposed in this study is not only simpler in structure and easier to fabricate but also exhibits higher absorptivity in the ultraviolet band. Additionally, it is characterized by stability at large incident angles and polarization independence.更多还原