The use of additive manufacturing for rapid prototyping of near-infrared and terahertz components provides seamless and error-free production. This article discusses the additive manufacturing and post-processing of axicons and their performance evaluation using attenuation and near-field-measurements based fundamental techniques. The axicons are manufactured using the materials cyclic olefin copolymer (TOPAS) and polymethyl methacrylate (PMMA), for their respective use in terahertz and near-infrared applications. The optical and terahertz components manufactured using traditional 3D-printing processes, e.g., fused filament fabrication or stereolithography apparatus exhibit high surface roughness in the range of 15 ± 2.5 µm, resulting in undesired propagation and scattering in the near infrared wavelengths. This research work proposes an economical post-processing technique for additively manufactured terahertz and near-infrared axicons for applications in multispectral characterization, e.g., bio-sensing. The authors used an enhanced method of dip-coating, which involves interval dipping and intermittent hardening to achieve better surface finish. An emphasis is placed on interval dipping and intermittent hardening, which lead to excellent transparency in case of additively-manufactured near-infrared axicons. The dip-coated samples exhibit surface roughness below 10 nm. With the use of heated resin material as the coating layer, due to reduced viscosity, the resin material distributes uniformly over the surface of the 3D-printed terahertz and near-infrared axicons. The authors also observed that the DOF length deviation between unprocessed and enhanced dip-coated axicons remains within the measurement error estimation from analytical calculations. In addition to the improved surface finish and transparency, the coatings are also closely matched in refractive index to the axicon material. Such post-processed axicons pave the way for producing a wide array of systems in the fields of communication, imaging, and bio-sensing.
Shrotri et al. (Tue,) studied this question.