Ultra-broadband composite terahertz absorber prediction based on K-nearest neighbor

Abstract

Broadband Terahertz (THz) absorbers with characteristics of thin thickness, wide absorption bandwidth, and tunability can play a significant role in the field of electromagnetic (EM) wave shielding. However, when designing absorbers, it is necessary to analyze a large number of variable structural parameters to obtain the desired broadband absorption spectrum. Therefore, it is difficult to obtain the optimal structure efficiently and accurately by artificially designing the broadband absorbers because a large amount of structural parameter optimization work is required. This study uses the K-nearest neighbor (KNN) algorithm to predict the design of composite multilayer terahertz absorber. The prediction obtained an ultra-large effective absorption bandwidth of 5.493 THz, with an absolute percentage error (APE) of only 0.69%. In addition, the proposed absorber has functions such as polarization insensitivity, stable performance for large incident angles, and absorption intensity tunability. This study will provide a new approach for the structure design of ultra-broadband terahertz absorbers in the future.